HOW YOU CAN USE
DATA ANALYTICS
TO CHANGE
GOVERNMENT
2 | A GOVLOOP GUIDE
2 | What Is Data Analytics?
7 | How to Make Data Analytics User-Friendly
11 | How Data Analytics Drives Real-Time and Forensic Analyses
15 | Accelerating Insights with Multi- Tiered Flash Storage
19 | The Future of Data Analytics in Government IT
23 | In the World of Data Analytics, Speed Is Everything
27 | 4 Keys to a Successful Data Analytics Program
31 | Data Analytics as an Asset
34 | 3 Reasons Privacy Should Be Top of Mind
35 | What You Can Do Now: How to Create a Culture for Data Analytics
37 | About Govloop & Acknowledgments
4 | Fighting the Opioid Epidemic
16 | Improving Human Capital Management
28 | Improving Mail Delivery
8 | Making Patent Data Searchable
12 | Tracking Medicare Fraud
24 | Predicting and Reducing Crime
20 | Empowering Cyber Analysts
32 | Enhancing IT Investment Tracking
CONTENTS
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 1
Everything about Dr. Farid Fata’s case seemed too egregious to be true.
When D�ZKLVWOHEORZHU�WLSSHG�R�WKH�'HSDUW-ment of Justice back in 2013�DERXW�'U��)DWDȇV�unprecedented medical malpractice, govern-
PHQW�RɝFLDOV�KDG�D�KDUG�WLPH�ZUDSSLQJ�WKHLU�KHDGV�DURXQG�WKH�FODLPV�ȃ�DW�ȴUVW��
How could a renowned Michigan oncologist
intentionally misdiagnose patients and then
inject them with chemotherapy treatments?
And how could he intentionally underdose
patients who did in fact have cancer?
In total, 553 patients and their loved ones
ZHUH�YLFWLPL]HG��6RPH�SDWLHQWV�GLHG�EHIRUH�VHHLQJ�'U��)DWD�VHQWHQFHG�WR����\HDUV�LQ�SULVRQ�RQ�-XO\�����������2WKHUV�ZHUH�OHIW�ZLWK�organ damage, missing teeth, nerve damage
DQG�SV\FKRORJLFDO�DQG�HPRWLRQDO�VFDUV�
Because of the nature of the harm, govern-
PHQW�RɝFLDOV�DUH�FDOOLQJ�'U��)DWDȇV�FDVH�WKH�PRVW�VHULRXV�IUDXG�FDVH�LQ�8�6��KLVWRU\��%ULQJ-
ing him to justice took a brave whistleblower,
government collaboration — and reliable
GDWD�DQDO\WLFV��
Ȋ2XU�LQYHVWLJDWRUV�SDUWQHUHG�ZLWK�WKH�)%Ζ��Ζ56�DQG�'R-��LQFOXGLQJ�WKH�8�6��$WWRUQH\ȇV�2ɝFH�LQ�WKH�(DVWHUQ�'LVWULFW�0LFKLJDQ�ȋ�VDLG�&DU\O�Brzymialkiewicz, Assistant Inspector General
DQG�&KLHI�'DWD�2ɝFHU�DW�WKH�'HSDUWPHQW�RI�+HDOWK�DQG�+XPDQ�6HUYLFHV�2ɝFH�RI�WKH�ΖQVSHFWRU�*HQHUDO��Ȋ7KH\�DOO�ZRUNHG�FROODERUD-
tively, nonstop, and pulled enough evidence
together to take quick action, including the
XVH�RI�GDWD�DQDO\WLFV�ȋ
When this case came in, the team checked
WKH�ZKLVWOHEORZHUȇV�OHDG�DJDLQVW�GLHUHQW�GDWD�sets, including claims data, Brzymialkiewicz
VDLG��7KH�GDWD�FRUURERUDWHG�WKH�ZKLVWOHEORZ-
HUȇV�FODLPV��EXW�LQLWLDOO\�LW�ZDV�VXFK�DQ�RXWOLHU�WKDW�RɝFLDOV�WKRXJKW�VRPHWKLQJ�ZDV�ZURQJ�ZLWK�WKH�GDWD��
Ȋ:H�WDON�DERXW�WZR�RXWFRPHV�IRU�WKLV�RɝFH��ΖWȇV�UHDOO\�DERXW�SURWHFWLQJ�EHQHȴFLDULHV��DV�ZHOO�DV�WKH�SURJUDPV�ȋ�VKH�VDLG��Ȋ6R�Ζ�WHOO�WKDW�story because, in terms of detecting fraud,
XVLQJ�GDWD�DQDO\WLFV�LV�D�KXJH�PLVVLRQ�ȋ�
Healthcare is one of the key areas in govern-
ment where data analytics is driving value,
EXW�LWȇV�FHUWDLQO\�QRW�WKH�RQO\�RQH��ΖQ�ΖQGLDQD��government employees are using data ana-
O\WLFV�WR�FRPEDW�WKH�VWDWHȇV�RSLRLG�HSLGHPLF��7KH�*HQHUDO�6HUYLFHV�$GPLQLVWUDWLRQ�LV�XVLQJ�human capital analytics to inform human
UHVRXUFHV�KLULQJ�DQG�UHFUXLWLQJ�GHFLVLRQV��$QG�there are countless other examples, several
RI�ZKLFK�\RXȇOO�ȴQG�LQ�WKLV�JXLGH��
:KHQ�XVHG�HHFWLYHO\��GDWD�DQDO\WLFV�FDQ�KHOS�WR�VDYH�OLYHV��LPSURYH�HɝFLHQFLHV��UHGXFH�costs, and help government deliver better cit-
L]HQ�VHUYLFHV��$QG�JRYHUQPHQW�LV�XVLQJ�WKHVH�FDSDELOLWLHV�PRUH�WKDQ�HYHU��VR�LWȇV�HVSHFLDOO\�important today for you to learn how and why
WKH\�IXQFWLRQ�
7R�VXSSRUW�\RX�RQ�\RXU�GDWD�DQDO\WLFV�MRXUQH\��ZHȇYH�SXW�WRJHWKHU�WKLV�*RY/RRS�UHVRXUFH�WR�KHOS�\RX�
• 'HȴQH�GDWD�DQDO\WLFV�DQG�XQGHUVWDQG�ZK\�LWȇV�D�ELJ�GHDO�LQ�JRYHUQPHQW
• Identify various use cases at the fed-HUDO��VWDWH�DQG�ORFDO�OHYHOV�DQG�WKHLU�XQGHUO\LQJ�NH\�VXFFHVV�IDFWRUV
• Understand the privacy concerns DVVRFLDWHG�ZLWK�GDWD�DQDO\WLFV
• *OHDQ�SUDFWLFDO�WLSV�IRU�EXLOGLQJ�D�FXO-WXUH�IRU�GDWD�DQDO\WLFV�DW�\RXU�DJHQF\
But before we dive into the meat of this
JXLGH��OHWȇV�HVWDEOLVK�WKH�EDVLFV�DERXW�GDWD�DQ-
DO\WLFV�ȃ�ZKDW�LW�LV��DQG�ZK\�\RX�VKRXOG�FDUH��
EXECUTIVE SUMMARY
2 | A GOVLOOP GUIDE
“
WHAT IS DATA ANALYTICS?Is it big data, data analytics, or just analytics?ΖI�\RXȇYH�HYHU�SRQGHUHG�WKLV�TXHVWLRQ��\RXȇUH�QRW�DORQH��2IWHQ�\RXȇOO�KHDU�WKHVH�ZRUGV�XVHG�LQWHUFKDQJHDEO\�RU�PHUJHG�LQWR�RQH�SKUDVH��ELJ�GDWD�DQDO\WLFV��ΖW�FDQ�EH�FRQIXVLQJ��ULJKW"�
%XW�GRQȇW�JHW�WRR�FDXJKW�XS�LQ�WKH�WHUPLQRORJ\�EHFDXVH�LWȇV�ERXQG�WR�FKDQJH�IURP�SHUVRQ�WR�SHUVRQ��)RU�WKH�VDNH�RI�WKLV�JXLGH��ZH�ZLOO�VWLFN�ZLWK�D�GHȴQLWLRQ�SURYLGHG�LQ�3UHVLGHQW�2EDPDȇV�Big 'DWD�DQG�3ULYDF\�:RUNLQJ�*URXS�5HYLHZ:
2QH�RI�WKH�UHDVRQV�\RXȇOO�RIWHQ�KHDU�ELJ�GDWD��GDWD�DQDO\WLFV�DQG�ELJ�GDWD�DQDO\WLFV�XVHG�LQ�WKH�VDPH�EUHDWK�LV�EHFDXVH�WKHUH�LV�D�ORW�RI�RYHUODS��:H�QHHG�ERWK�WKH�WRROV�WR�FROOHFW�DQG�PDQDJH�ODUJHV�YROXPHV�RI�YDU\LQJ�GDWD��DV�ZHOO�DV�WKH�WRROV�WR�PDNH�VHQVH�of the data and turn that insight into action.
:KHQ�\RX�KHDU�WKH�ZRUG�ȊELJȋ�FRXSOHG�ZLWK�ȊDQDO\WLFV�ȋ�FRQVLGHU�WKHVH�WZR�WKLQJV��ΖI�WKH�GDWD�WKDWȇV�FROOHFWHG�DQG�SURFHVVHG�LV�ELJ�LQ�TXDQWLW\�DQG�YDULHW\��WKHQ�LWȇV�OLNHO\�WKH�DQDO\VLV�DQG�GDWD�FRUUHODWLRQV�ZLOO�EH�ELJ�WRR��
%XW�GDWD�DQDO\WLFV�LV�DERXW�PRUH�WKDQ�DOJRULWKPV��ΖWȇV�DOVR�DERXW�WKH�SK\VLFDO�SODWIRUPV�DJHQFLHV�QHHG�WR�VWRUH�DQG�DQDO\]H�WKHLU�GDWD��7R�PHHW�WKRVH�QHHGV��PDQ\�DJHQFLHV�DUH�WXUQLQJ�WR�FORXG�FRPSXWLQJ��ZKLOH�DOVR�EDODQFLQJ�WKH�QHHG�IRU�VHFXULW\�DQG�SULYD-F\��:HȇOO�JHW�LQWR�WKH�PDMRU�SULYDF\�LVVXHV�ODWHU�LQ�WKH�JXLGH��
Analytics is what makes big data come alive.
Without analytics, big datasets could be stored, and they could
be retrieved, wholly or selectively. But what comes out would
be exactly what went in. Analytics, comprising a number of dif-
ferent computational technologies, is what fuels the big-data
revolution. Analytics is what creates the new value in big data-
sets, vastly more than the sum of the values of the parts.”
DESCRIPTIVEANALYTICS
What happened?
DIAGNOSTICANALYTICS
Why did it happen?
PREDICTIVEANALYTICS
What will happen?
PRESCRIPTIVEANALYTICS
How can we make it happen?
D I F F I C U L T Y
I N F O R M A T I O N
O P T I M I Z A T I O N
VALUE
Using Google Analytics to track website traffic, such as page views and number of visitors
Mining data to determine what caused a spike in Web traffic over the past month
Using algorithms and data to predict where burglaries will most likely occur in a city
Rerouting travel, based on traffic predictions and pedestrians’ actions (think self-driving cars)
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 3
4 TYPES OF DATA ANALYTICS
ANALYTIC VALUE ESCALATOR
APPLY IT NOW
8QWLO�WKLV�SRLQW��ZHȇYH�XVHG�GDWD�DQDO\WLFV�DV�D�EODQNHW�WHUP�WR�GH-VFULEH�KRZ�DJHQFLHV�JOHDQ�LQVLJKW�IURP�WKHLU�GDWD�WR�VROYH�UHDO�SURE-OHPV�DQG�FDUU\�RXW�WKHLU�PLVVLRQV��%XW�WKHUH�DUH�IRXU�GLVWLQFW�W\SHV�RI�GDWD�DQDO\WLFV�WKDW�KHOS�DJHQFLHV�DQVZHU�WKH�IROORZLQJ�TXHVWLRQV��
:HȇYH�SXOOHG�WRJHWKHU�VRPH�LQVLJKWV�IURP�*DUWQHU�WR�KHOS�\RX�XQGHUVWDQG�WKH�IRXU�W\SHV�RI�DQDO\WLFV�DQG�YLVXDOL]H�WKH�ULVLQJ�GHJUHH�RI�GLɝFXOW\�DQG�YDOXH�WLHG�WR�HDFK�RI�WKHP��<RXȇOO�QRWLFH�WKHVH�WHUPV��GHVFULSWLYH��GLDJQRVWLF��SUHGLFWLYH��DQG�SUHVFULSWLYH��PHQWLRQHG�WKURXJKRXW�WKH�JXLGH��HVSHFLDOO\�LQ�RXU�FDVH�VWXGLHV�
7DNH�D�PLQXWH�DQG�WKLQN�DERXW�RQH�RI�\RXU�RUJDQL]DWLRQȇV�ELJJHVW�SDLQ�SRLQWV��RU�SHUKDSV�D�FKDOOHQJH�\RX�IDFH�LQ�\RXU�HYHU\GD\�ZRUN�OLIH��0D\EH�WKHUHȇV�D�VSHFLȴF�SURMHFW�\RXȇUH�ZRUNLQJ�RQ�ULJKW�QRZ��7KLQN�DERXW�WKH�W\SH�RI�LQVLJKWV�RU�HɝFLHQFLHV�\RX�FRXOG�JDLQ�LI�\RX�XQGHUVWRRG�ZKDW�KDSSHQHG��ZK\�LW�KDSSHQHG��ZKDW�ZLOO�OLNHO\�KDSSHQ��DQG�ZKDW�WR�GR�DERXW�LW��
1What happened? 2Why did it happen?
3What will happen? 4How can we make it happen?
DESCRIPTIVEANALYTICS
What happened?
DIAGNOSTICANALYTICS
Why did it happen?
PREDICTIVEANALYTICS
What will happen?
PRESCRIPTIVEANALYTICS
How can we make it happen?
D I F F I C U L T Y
I N F O R M A T I O N
O P T I M I Z A T I O N
VALUE
Using Google Analytics to track website traffic, such as page views and number of visitors
Mining data to determine what caused a spike in Web traffic over the past month
Using algorithms and data to predict where burglaries will most likely occur in a city
Rerouting travel, based on traffic predictions and pedestrians’ actions (think self-driving cars)
4 | A GOVLOOP GUIDE
The State of IndianaAn interview with Dave Matusoff & Joshua MartinDirector for the Indiana Management & Performance Hub; MPH Chief of Staff
In Indiana, news stories about Hoosiers who are addicted to opioid are
GULYLQJ�PRUH�RSHQ�FRQYHUVDWLRQ�DERXW�WKH�VWDWHȇV�JURZLQJ�GUXJ�HSL-GHPLF��7KH�VWRULHV�DUH�DERXW�SHRSOH�OLNH�&KHU\O�'LFNHQ��ZKR��according
WR�WKH�ΖQGLDQDSROLV�6WDU, “found her son unconscious in a heroin-in-
GXFHG�KD]H�ȋ�DQG�-RDQ�0RRQ��ȊZKRVH�JUDQGVRQ�KDV�EHHQ�DGGLFWHG�WR�GUXJV�VLQFH������ȋ
7KHVH�KXUWLQJ�IDPLOLHV�DUH�WKH�IDFH�RI�D�ODUJHU�SUREOHP�WKDW�GHPDQGV�PRUH�DQVZHUV��PRUH�LQVLJKW�DQG��LQHYLWDEO\��PRUH�PHDQLQJIXO�GDWD��Although personal stories humanize the drug epidemic, stories alone
DUH�QRW�WKH�VROXWLRQ��/DZ�HQIRUFHPHQW�RɝFLDOV��SXEOLF�KHDOWK�DJHQFLHV��lawmakers, and the community need sound data to shape their collec-
WLYH�UHVSRQVH�WR�WKH�SUREOHP��
%XW�ZKHQ�WKDW�GDWD�LV�VWRUHG�LQ�GLHUHQW�V\VWHPV�DW�GLHUHQW�DJHQFLHV��PDNLQJ�XVH�RI�LW�UHTXLUHV�KDUG�ZRUN�DQG�FROODERUDWLRQ��
6SXUUHG�E\�VHULRXV�FRQFHUQV�DERXW�WKH�ULVH�RI�RSLRLG�XVH�VWDWHZLGH�ȃ�DQG�D�IDVFLQDWLRQ�ZLWK�GDWD�ȃ�-RVKXD�0DUWLQ��&KLHI�RI�6WD�IRU�WKH�ΖQGLDQD�0DQDJHPHQW�DQG�3HUIRUPDQFH�+XE��03+���GHFLGHG�WR�DFW��
ΖW�ZDV�LQ�$XJXVW������WKDW�0DUWLQ�QRWLFHG�D�ULVH�LQ�PHGLD�FRYHUDJH�RI�KHURLQ�UHODWHG�LQFLGHQWV��7KDW�ZDV�DERXW�WKH�VDPH�WLPH�03+�GHSOR\HG�a new business intelligence tool to help its agency customers make
GDWD�LQIRUPHG�GHFLVLRQV��
7KH�FXOPLQDWLRQ�RI�WKHVH�HYHQWV�JRW�0DUWLQ�WKLQNLQJ��:RXOGQȇW�LW�EH�LQ-
teresting if the state could use data to better understand the changing
nature of the drug problem?
Historically, decision-makers have had to rely on whatever information
was available to make the tough choices, but what they had access to
ZDV�QRW�DOZD\V�WKH�EHVW�LQIRUPDWLRQ��6RPHWLPHV�WKH�SHRSOH�ZKR�ZHUH�DVNLQJ�WKH�TXHVWLRQV�GLGQȇW�KDYH�DFFHVV�WR�WKH�ULJKW�GDWD��DQG�QHLWKHU�GLG�WKH�SHRSOH�ZULWLQJ�SROLFLHV��
Ȋ7KHUHȇV�D�UHDO�KXPDQ�HOHPHQW�WR�XVLQJ�WKH�GDWD�WR�LQIRUP�WKH�SROLF\�GHFLVLRQ�PDNLQJ�DURXQG�WKHVH�SUREOHPV�ȋ�VDLG�'DYH�0DWXVR��([HFX-
WLYH�'LUHFWRU�RI�03+��
THE CHALLENGE: Fighting the Opioid Epidemic
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 5
1 2 3
It has only been a few months since MPH stood up its drug data dash-
board internally and began enlisting the help of other agencies — the
GDWD�RZQHUV��7KH�SURMHFW�LV�VWLOO�LQ�LWV�LQIDQF\��EXW�WKDW�KDVQȇW�GHWUDFWHG�IURP�WKH�UHDO�ZRUN�XQGHUZD\��
7KH�03+�WHDP�KDV�ZRUNHG�FORVHO\�ZLWK�ΖQGLDQDȇV�SROLFH�IRUHQVLF�laboratories to track and visualize the types of illegal substances that
RɝFHUV�DUH�ȴQGLQJ�RQ�WKH�VWUHHW��DQG�KRZ�RIWHQ�WKH\ȇUH�ȴQGLQJ�WKHP��7KH�PRVW�UHFHQW�GDWD��WR�1RYHPEHU�������VKRZHG�D�GUDPDWLF�ULVH�in heroin submissions to the lab but a steady decline in other opioid
VXEPLVVLRQV��QDPHO\�SUHVFULSWLRQ�GUXJV��
0DWXVR�SRLQWHG�WR�D�2013 state law aimed at tackling the opioid
epidemic as a contributing factor in the decline of prescription opioid
DEXVH��7KH�ODZ�ZDV�SDVVHG�WR�WLJKWHQ�SUHVFULELQJ�SUDFWLFHV��FORVH�GRZQ�VR�FDOOHG�ȊSLOO�PLOOVȋ�ZKHUH�SK\VLFLDQV�DQG�RWKHU�SURYLGHUV�LOOHJDOO\�prescribe and issue narcotics, and reduce opioid abuse and ultimately
GHDWKV��7KH�*HQHUDO�$VVHPEO\ȇV�LQWHQWLRQV�ZHUH�JRRG��EXW�WKH�GDWD�VKRZV�XQLQWHQGHG�FRQVHTXHQFHV��
“When they passed that legislation, it was harder for addicts to get
SUHVFULSWLRQ�GUXJV�ȋ�0DWXVR�VDLG��Ȋ<RX�VDZ�D�GUDPDWLF�GHFOLQH�LQ�the number of prescription narcotics submitted to the forensics lab,
EXW�WKHQ�\RX�VDZ�D�UHDOO\�GUDPDWLF�ULVH�LQ�WKH�KHURLQ�XVH��6R�\RX�FDQ�begin to look at — in a very real-time way — the unintended conse-
TXHQFHV�RI�SROLF\�ȋ�
ΖQ�1RYHPEHU�������WKH�03+�WHDP�VWDUWHG�IRUPDOL]LQJ�WKH�GDVKERDUG�
and soliciting data and input from law enforcement and public health
DJHQFLHV��8VLQJ�WKH�EXVLQHVV�LQWHOOLJHQFH�WRRO��03+�FDQ�YLVXDOL]H�GDWD�by creating geospatial maps, color-coordinated line graphs with trend
data and even a time-lapse map that shows how the drug epidemic
KDV�HYROYHG�RYHU�WLPH��
03+�LVQȇW�MXVW�FROOHFWLQJ�GDWD�RQ�ODERUDWRU\�WHVWHG�VXEVWDQFHV��7KH�team is also looking at opioid-related deaths and geospatial data for
FHUWLȴHG�WUHDWPHQW�SURYLGHUV�DFURVV�WKH�VWDWH��
7KLV�LV�WKH�ȴUVW�WLPH�DOO�WKHVH�GDWD�VHWV�KDYH�EHHQ�FRPELQHG�WR�WHOO�D�FRPSOHWH�VWRU\��0DUWLQ�VDLG��$�TXLFN�ORRN�DW�WKHVH�YLVXDO�PDSV�PDNHV�FOHDU�WKDW�WKH�GUXJ�SUREOHPV�YDU\�E\�FRXQW\��
2QH�FRORU�FRGHG�PDS�LQ�SDUWLFXODU�VKRZV�WKH�QXPEHU�RI�SHU�FDSLWD��RSLRLG�UHODWHG�GHDWKV�E\�FRXQW\��6RXWKHDVWHUQ�ΖQGLDQD�KDV�EHHQ�KLW�E\�WKLV�HSLGHPLF�HVSHFLDOO\�KDUG��ΖQ�2FWREHU�������WKH�ΖQGLDQDSROLV�6WDU�UH-
ported�ȊPRUH�WKDQ�����SHRSOH�LQ�RU�QHDU�6FRWW�&RXQW\�ZHUH�GLDJQRVHG�ZLWK�+Ζ9�DIWHU�VKDULQJ�QHHGOHV�WR�LQMHFW�SUHVFULSWLRQ�SDLQ�PHGLFDWLRQ�ȋ
03+�LV�H[SORULQJ�WKH�UROH�WKDW�GDWD�ZLOO�SOD\�LQ�FRPEDWLQJ�WKH�VWDWHȇV�GHYDVWDWLQJ�GUXJ�SUREOHPV��7KH�WHDP�DOVR�KDV�LWV�VLJKWV�VHW�RQ�SURYLGLQJ�VLPLODU�GDWD�DQDO\WLFV�VHUYLFHV�WR�RWKHU�VWDWHV�IDFLQJ�WKHLU�RZQ�LVVXHV��
Ȋ7KLV�LV�VWLOO�D�ZRUN�LQ�SURJUHVV��EXW�ZKDW�ZHȇUH�WU\LQJ�WR�GR�KHUH�LV�WDNH�WKHVH�PRXQWDLQV�RI�GDWD�WKDW�DUH�LQ�GLHUHQW�SODFHV�DQG�WXUQ�GDWD�LQWR�information, turn that information into knowledge, and hopefully turn
WKDW�NQRZOHGJH�LQWR�SROLF\�ȋ�0DUWLQ�VDLG��
SUCCESS FACTORS
THE SOLUTION
Think through your strategy. Create a proof of concept, or partner with other organizations that have already made
investments.
Rethink how you can use your data in new and meaningful ways.
Solicit the help of administration officials to break down data-sharing
silos among agencies.
“What we’re trying to do here is take these mountains of data that are in different places and turn data into information, turn that information into knowledge, and hopefully turn that knowledge into policy.”
https://www.oracle.com/solutions/business-analytics/resources.html
LEARN MORE
DATA ANALYTICS
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 7
How to Make Data Analytics User-FriendlyAn interview with Rob LindsleyDirector of Data Analytics, Oracle Public Sector
INDUSTRY SPOTLIGHT
By now, data analytics has become one of the
ELJJHVW�WUHQGV�LQ�JRYHUQPHQW��$JHQFLHV�QHHG�LW��FRPSDQLHV�DUH�RHULQJ�LW��DQG�VROXWLRQV�DUH�GULYHQ�E\�LW��2QH�RI�WKH�PRVW�LPSRUW-ant aspects of data analytics is its ability to
integrate with existing systems, in terms of
data sources, applications, infrastructure, as
ZHOO�DV�VHFXULW\��%XW�LW�FDQ�WDNH�D�ORW�RI�ZRUN�WR�UHDFK�WKDW�OHYHO�RI�LQWHJUDWLRQ��
5RE�/LQGVOH\��'LUHFWRU�RI�'DWD�$QDO\WLFV�IRU�2UDFOH�3XEOLF�6HFWRU��VSRNH�ZLWK�*RY/RRS�about ways that organizations can work to
make data analytics as user-friendly and
ȵH[LEOH�DV�SRVVLEOH��2UDFOHȇV�RZQ�VWUDWHJ\�LV�focused on providing organizations with a va-
riety of cloud options — allowing for self-ser-
vice analytics and data visualization — and
presenting big data and predictive analysis
ZLWKRXW�D�VWHHS�OHDUQLQJ�FXUYH��6LPSO\�SXW��the strategy is to give customers “access to
the data they need paired with the analytics
WKH\�ZDQW�ȋ�DFFRUGLQJ�WR�/LQGVOH\��
7KH�ZD\�JRYHUQPHQW�DJHQFLHV�SURYLGH�DQDO\WLFV�WR�WKHLU�XVHUV�LV�DOVR�FKDQJLQJ��7KH\�can run analytic solutions in their own data-
FHQWHUV��FRPSDQLHV�WKDW�RHU�WKH�DQDO\WLFV�VHUYLFHV�FDQ�GHOLYHU�WKHP�DV�FORXG�RHULQJV��or agencies can have third parties run their
GDWDFHQWHUV��5HJDUGOHVV��WKH�WRROV�RU�VHUYLFHV�should be easily available to consumers both
LQVLGH�DQG�RXWVLGH�WKH�RUJDQL]DWLRQ��
/LQGVOH\�GHVFULEHG�2UDFOHȇV�UROH�LQ�JRYHUQ-
ment agencies as running the gamut from
FRUH�EXVLQHVV�RSHUDWLRQV�OLNH�ȴQDQFH�DQG�human resources to mission support func-
WLRQDOLW\�DQG�EH\RQG��Ȋ2UDFOH�KDV�D�VWURQJ�presence in the central operations of many
government agencies, but we also have a lot
of customers using analytics to accomplish
WKHLU�DJHQF\ȇV�FRUH�PLVVLRQ�LQ�D�YDULHW\�RI�DU-eas, including transportation, a child services
DQG�LQ�GHIHQVH�RUJDQL]DWLRQV��7KH\ȇUH�XVLQJ�analytics to meet their objectives quickly and
PRUH�HɝFLHQWO\�ȋ
2QH�RI�WKH�KDOOPDUNV�RI�2UDFOHȇV�DQDO\W-LFV�SODWIRUP�LV�WKDW�LWȇV�RHUHG�LQ�ERWK�DQ�on-premise deployment model or in the
FORXG��8VHUV�FDQ�VHOHFW�ZKLFK�RSWLRQ�EHVW�VXLWV�WKHLU�QHHGV��)RU�WKH�FORXG�RSWLRQ��XVHUV�FDQ�FKRRVH�D�SXEOLF��SULYDWH�RU�K\EULG�PRGHO��)OH[LELOLW\�LV�NH\��EHFDXVH�XVHUV�FDQ�HDVLO\�change between on-premise deployment
and cloud deployment as projects develop
DQG�DJHQFLHVȇ�QHHGV�HYROYH�RYHU�WLPH��)RU�example, an agency might test a new system
in the cloud and then move it to an on-prem-
LVHV�GDWDFHQWHU�IRU�SURGXFWLRQ�GHSOR\PHQW��2U�WKH\�PLJKW�XVH�FORXG�LQIUDVWUXFWXUH�IRU�SXEOLF�IDFLQJ�ZHEVLWHV�DQG�RHU�LPSURYHG�data transparency
A variety of data sources are essential to
LPSURYH�ȵH[LELOLW\�IRU�DQDO\WLFV�SODWIRUPV��7UD-
GLWLRQDO�GDWD�DQDO\WLFV�HPSOR\V�ȊRɝFLDOȋ�GDWD�warehouses or other standard data models,
but increasingly government employees need
to integrate their own data into the analysis,
whether through an Excel spreadsheet or a
FXUUHQW�URVWHU�RI�JRYHUQPHQW�FRQWUDFWRUV��2UDFOH�GHOLYHUV�VHOI�VHUYLFH�FDSDELOLWLHV�WKDW�allow end-users to incorporate a variety of
GDWD�VRXUFHV�LQWR�WKHLU�DQDO\WLFV�
Agencies are also pushing to instill a culture
RI�GDWD�GULYHQ�GHFLVLRQ�PDNLQJ���%RWK�ELJ�data analytics and predictive analysis have
VLJQLȴFDQW�SRWHQWLDO�IRU�HPSOR\HHV�ZKHQ�they can be used without a steep learning
FXUYH��7KH�JRYHUQPHQW�KDV�YDVW�DPRXQWV�RI�information on the services it provides, and it
also has access to information about external
HYHQWV��$Q�H[DPSOH�RI�DQ�DJHQF\�VXFFHVVIXOO\�harnessing the power of big data analytics is
WKH�'HSDUWPHQW�RI�7UDQVSRUWDWLRQ��'27��LQ�D�PLG�ZHVWHUQ�VWDWH��%\�FRPELQLQJ�GDWD�RQ�WUDI-ȴF�SDWWHUQV��URDG�FRQGLWLRQV��VDIHW\�IHDWXUHV��and the weather, this agency can optimize the
travel experience, make best use of limited
resources and increase the safety of drivers
DFURVV�WKH�VWDWH��
2UDFOH�KDV�DOVR�KHOSHG�DJHQFLHV�XVH�H[LVWLQJ�data to improve current projects and predict
IXWXUH�WUHQGV��2QH�SURMHFW�LQYROYHG�WKH�8�6��Army, which for more than a decade has
ZRUNHG�ZLWK�2UDFOH�WR�KHOS�LGHQWLI\�ZKLFK�XQLWV�DUH�EHVW�VXLWHG�IRU�SDUWLFXODU�PLVVLRQV��7KH�SURFHVV�DJJUHJDWHV�QXPHURXV�GDWDVHWV��LQFOXGLQJ�WKRVH�IRU�VWDɝQJ��WUDLQLQJ�OHYHOV�DQG�H[LVWLQJ�UHVRXUFHV��7KH�$UP\�VWDUWHG�with straightforward metrics to evaluate unit
readiness and then moved toward descriptive
DQDO\VLV��XVLQJ�LQIRUPDWLRQ�DERXW�VSHFLȴF�FDSDELOLWLHV�XQLTXH�WR�WKH�XQLW��0RUH�UHFHQWO\��the Army is using predictive analysis to deter-
mine what types of units will be required in
IXWXUH�VFHQDULRV���
/LQGVOH\�QRWHV�WKHVH�H[DPSOHV�KLJKOLJKW�the democratization of data access, the
usability of modern analytics tools, and easy
LPSOHPHQWDWLRQ��2UDFOHȇV�VWUHQJWK�LV�VROYLQJ�FRPSOH[�WHFKQRORJ\�FKDOOHQJHV��KH�VD\V��Ȋ)RU�JRYHUQPHQW��WKLV�PHDQV�RXU�DQDO\WLFV�initiatives can help agencies address a variety
RI�UHTXLUHPHQWV�Ȃ�ZKHWKHU�LWȇV�PDVVLYH�GDWD�VHWV��VWULQJHQW�VHFXULW\��RU�VFDODELOLW\��7KHVH�examples remind us what can be achieved
when we can seamlessly incorporate data
DQDO\WLFV�LQWR�RXU�SUREOHP�VROYLQJ�VWUDWHJLHV�ȋ�
8 | A GOVLOOP GUIDE
We can all agree that managing a growing database of more than
����PLOOLRQ�UHFRUGV�LV�QR�VPDOO�IHDW��:KDW�PDNHV�WKLV�XQGHUWDNLQJ�HYHQ�PRUH�FKDOOHQJLQJ�LV�ȴJXULQJ�RXW�KRZ�EHVW�WR�PDNH�GHFDGHV�ROG�UHFRUGV�HOHFWURQLFDOO\�VHDUFKDEOH�DQG�DFFHVVLEOH�WR�WKH�SXEOLF��
)RU�WKH�8�6��3DWHQW�DQG�7UDGHPDUN�2ɝFH��86372���JHWWLQJ�WKLV�ULJKW�ZDV�QRW�RQO\�DQ�DJHQF\�LQLWLDWLYH�EXW�DOVR�D�:KLWH�+RXVH�SULRULW\��
)URP�WKH�SXEOLFȇV�SHUVSHFWLYH��WKHUHȇV�D�JUHDW�GHPDQG�IRU�WKH�EXON�DVVLJQPHQW�GDWD�WKDW�86372�PDNHV�DYDLODEOH�YLD�LWV�online search tool,
DSSURSULDWHO\�QDPHG�$VVLJQPHQW�6HDUFK��Ȋ:H�JHW�EHWZHHQ�������DQG�������VHDUFKHV�D�GD\�ȋ�VDLG�7HG�3DUU��'LUHFWRU�RI�WKH�3XEOLF�ΖQIRUPDWLRQ�6HUYLFHV�*URXS�DW�86372��Ȋ7KHUH�DUH�DERXW�������GLHUHQW�XVHUV�HDFK�ZHHN�ZKR�XVH�WKH�WRRO�ȋ
7KH�WRRO�DOORZV�XVHUV�WR�VHDUFK�IRU�WKH�ȊDVVLJQPHQWV�ȋ�RU�OHJDO�SDSHU-work, that document the transfer of a patent from a previous owner to
WKH�QHZ�RZQHU��ΖW�FRXOG�EH�D�ZLOO�RU�D�FRQWUDFW�WKDW�LQFOXGHV�WKH�QDPH�of the previous owner, the new owner, their addresses and the type of
WUDQVIHU�EHLQJ�PDGH��
Ȋ:HȇYH�KDG�WKH�VHDUFK�WRRO�RQ�WKH�:HE�IRU�TXLWH�D�ZKLOH��EXW�LW�ZDV�UDWKHU�EDVLF�ȋ�3DUU�VDLG��ȊΖW�KDG�RQO\�D�IHZ�RI�WKH�GDWD�ȴHOGV�WKDW�ZH�have in our database open for searching, and it really was just a data-
EDVH�ORRNXS�WRRO��ΖW�GLGQȇW�KDYH�DQ\�NLQG�RI�DGYDQFHG�IXQFWLRQDOLW\��RU�HYHQ�ZKDW�\RX�ZRXOG�FRQVLGHU�EDVLF�IXQFWLRQDOLW\�QRZDGD\V�ȋ
)RU�H[DPSOH��LQ�WKH�ROG�V\VWHP��D�VHDUFK�IRU�D�FRPPRQ�DVVLJQHH�FRXOG�\LHOG�KXQGUHGV�RI�SDJHV�RI�UHVXOWV�DQG�UHTXLUH�KRXUV�RI�DQDO\VLV��7R�see all the assignments from Motorola to Google using the old system,
D�XVHU�ZRXOG�KDYH�WR�VHDUFK�IRU�0RWRUROD�DV�WKH�DVVLJQRU��7KDW�VHDUFK�alone produces 685 assignments; and each one would have to be
RSHQHG�WR�VHH�LI�LW�ZDV�DVVLJQHG�WR�*RRJOH�
7KHUH�ZHUH�VHDUFKHV�WKDW�HYHQ�86372�HPSOR\HHV�FRXOGQȇW�GR�ZLWKRXW�HQJDJLQJ�D�GHYHORSHU�WR�FRQFRFW�D�VSHFLȴF�TXHU\�IRU�WKHP��
THE CHALLENGE: Making Patent Data Searchable
U.S. Patent & Trademark OfficeAn interview with Ted ParrDirector of the Public Information Services Group within the Office of the CIO at USPTO
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 9
1 2 3
7RGD\��WKRVH�VHDUFKHV�DUH�IDU�OHVV�FXPEHUVRPH�IRU�ERWK�86372�HP-
SOR\HHV�DQG�WKH�SXEOLF��WKDQNV�WR�PRUH�UREXVW�DQDO\WLFV��
ȊΖWȇV�QRW�VR�PXFK�WKDW�ZH�SXW�LQ�SODFH�D�VSHFLȴF�DQDO\WLFV�FDSDELOLW\��EXW�ZH�HQDEOHG�LW�ȋ�3DUU�VDLG��Ȋ7KH�QHZ�V\VWHP�XVHV�6ROU�RSHQ�VRXUFH�software to provide indexed searching rather than the slower, more
OLPLWHG�GDWDEDVH�ORRNXSV�ZH�SUHYLRXVO\�XVHG�ȋ
ΖQ�DQ�HRUW�WR�PDNH�GDWD�PRUH�WUDQVSDUHQW�WR�WKH�SXEOLF�DQG�HDVLHU�IRU�SHRSOH�WR�ȴQG��DQG�WR�HQKDQFH�LQQRYDWLRQ��IDFLOLWDWH�WHFKQRORJ\�transfers and reduce abusive patent litigation and lawsuits stemming
IURP�SDWHQW�LQIULQJHPHQW��86372�IRFXVHG�LWV�DWWHQWLRQ�RQ�HQKDQFLQJ�WKH�VHDUFK�IXQFWLRQDOLW\�RI�SDWHQW�DVVLJQPHQWV��
Ȋ7KH�QHZ�V\VWHP�DOORZV�VHDUFKLQJ�E\�VHYHUDO�DGGLWLRQDO�GDWD�ȴHOGV��LQ-
cluding execution date, conveyance type, correspondent and assignee
address, as well as allowing wild card and Boolean searches for greater
SUHFLVLRQ�ȋ�3DUU�VDLG��Ȋ7KH�QHW�UHVXOW�LV�WKDW�FXVWRPHUV�FDQ�EH�YHU\�SUHFLVH�ZLWK�WKHLU�VHDUFKHV�DQG�FDQ�ȴOWHU�DQG�DQDO\]H�UHVXOWV�LQ�PDQ\�GLHUHQW�ZD\V�ȋ
5HPHPEHU�WKH�*RRJOH�DQG�0RWRUROD�H[DPSOH�ZH�JDYH�HDUOLHU"�8VLQJ�the new system, users simply enter both names in their respective
ȴHOGV��DQG�ZLWKLQ�D�IHZ�VHFRQGV�WKH\�ZLOO�VHH�WKH�����DVVLJQPHQWV�from Motorola to Google, rather than having to sift through more than
WZLFH�WKDW�QXPEHU�RI�OHVV�SUHFLVH�VHDUFK�UHVXOWV��
“In addition, the new tool will have a true data export feature in its next
UHOHDVH�LQ�D�IHZ�ZHHNV�ȋ�3DUU�QRWHG��
0DMRU�XSJUDGHV�ZHUH�UROOHG�RXW�LQ�'HFHPEHU�������LQFOXGLQJ�DQ�DSSOL-cation programming interface that allows users to build their own tools
WR�DFFHVV�86372ȇV�DVVLJQPHQW�V\VWHP�DQG�WKH�GDWD�EHKLQG�LW��8VHUV�can also now view the underlying legal documents related to their
VHDUFKHV��UDWKHU�WKDQ�MXVW�D�FRYHUVKHHW��
7KH�LQLWLDO�SURMHFW�NLFNHG�R�LQ������EHWZHHQ�$XJXVW�DQG�6HSWHPEHU��As a major transparency initiative, it drew interest from the White
+RXVH��ZKLFK�JDYH�LPSHWXV�IRU�D�TXLFN�FRPSOHWLRQ��
Ȋ6R�ZH�UHDOO\�GLG�IDVW�WUDFN�LW�WR�JHW�LW�RXW�E\�'HFHPEHU�ȋ�3DUU�VDLG��
7KDW�WLPHOLQH�LV�UDUH�LQ�JRYHUQPHQW��EXW�WKHUH�ZHUH�D�QXPEHU�RI�IDFWRUV�NH\�WR�WKH�SURMHFWȇV�VXFFHVV��:KLWH�+RXVH�RɝFLDOV�DQG�SROLF\-PDNHUV�ZHUH�LQYROYHG�HDUO\�RQ��DQG�WKH\�VXSSRUWHG�WKH�SURMHFW��3DUU�and his team used agile development to add user capabilities in rapid
LQFUHPHQWV��WKH\�ZRUNHG�HDUO\�DQG�RIWHQ�ZLWK�SHRSOH�ZKR�ȴOH�DVVLJQ-
ments and other users to incorporate their feedback; and they were
DEOH�WR�XVH�DQ�H[LVWLQJ�FRQWUDFW�UDWKHU�WKDQ�FUHDWLQJ�D�QHZ�RQH��
Ȋ7KDW�ZDV�D�ELJ�KHOS�WR�XV�ȋ�3DUU�VDLG��Ȋ7KHUH�ZDV�QR�ZD\�ZH�ZHUH�JRLQJ�WR�PDNH�RXU�-DQ����GHDGOLQH�ZLWKRXW�LW�ȋ
SUCCESS FACTORS
THE SOLUTION
Get high-level officials involved early on.
Let the agency’s business units share in the decision-making.
Explore agile development & actively participate in daily scrum
& project meetings.
6.7 6k 3k 4M I L L I O N
R E C O R D SU S E R S
P E R W E E KS E A R C H E S
P E R D A YM O N T H
T I M E L I N E
“The net result is that customers can be very precise with their searches and can filter and analyze results in many different ways.”
10 | A GOVLOOP GUIDE
Stop Cyber Threats In Their Tracks
Put Big Data Analytics to Work in Your Cyber Environment. ViON’s DataAdapt S2000 Cyber Secure analyzes large volumes of sensor data in near-real time to:
• Identify and thwart cyber attacks
• Eliminate insider threat vulnerabilities
• Provide actionable predictions for Advanced Persistent Threats (APTs)
LEARN MOREwww.vion.com/Solutions/Cyber-Security.aspx
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 11
How Data Analytics Drives Real-Time and Forensic AnalysesAn interview with Billy HillCustomer Solution Architect for Big Data Analytics at ViON
INDUSTRY SPOTLIGHT
7KH�GDWD�DQDO\WLFV�ODQGVFDSH�LV�UDSLGO\�HYROY-LQJ�WKDQNV�WR�PRUH�SRZHUIXO�DQG�DRUGDEOH�WRROV�IRU�GDWD�JDWKHULQJ�DQG�SURFHVVLQJ��
0HPRU\��&38V��DQG�IDVW�VWRUDJH�DUH�JHWWLQJ�cheaper as the manufacturing processes
DUH�LPSURYHG���(DFK�JHQHUDWLRQ�LV�EHFRPLQJ�more familiar with the ecosystem of analytics
and data storage technology, said Billy Hill,
&XVWRPHU�6ROXWLRQ�$UFKLWHFW�IRU�%LJ�'DWD�$QDO\WLFV�DW�9L21��7KLV�FRPELQDWLRQ�RI�DRUG-
ability and usability bring analytics tools within
reach for government agencies, businesses
DQG�HYHQ�FRQVXPHUV�
“Although the technology has been around
IRU�\HDUV��LWȇV�QRZ�DFFHVVLEOH�WR�D�ZLGHU�DXGL-HQFH�ȋ�+LOO�VDLG��
3HRSOH�DUH�XVLQJ�PRUH�&38V�DQG�PHPRU\�LQ�concert, and these technologies are start-
LQJ�WR�EH�LQFRUSRUDWHG�LQWR�WKH�ZRUNȵRZ�of researchers and mathematicians, who
XOWLPDWHO\�GHYHORS�WKH�DOJRULWKPV�XVHG�WR�ȴQG�SDWWHUQV�LQ�WKH�GDWD��$GYDQFHV�LQ�KDUGZDUH�are helping to drive changes in the types of
DOJRULWKPV�WKDW�DUH�GHYHORSHG��$W�WKH�VDPH�time, the algorithms are also dictating how
WKH�KDUGZDUH�LV�XVHG��ȊΖWȇV�D�V\QHUJLVWLF�GHYHO-RSPHQW�ȋ�+LOO�VDLG��
As the technology advances and provides
better insights, the demand for data analytics
and the outcomes they produce continues to
LQFUHDVH��Ȋ/HDGHUV�QRZDGD\V�DUH�H[SHFWLQJ�WR�JHW�PRUH�YDOXH�RXW�RI�WKHLU�GDWD�ȋ�+LOO�VDLG��
7KDWȇV�ZK\�DJHQFLHV�DUH�XVLQJ�GLHUHQW�W\SHV�of analytics, such as predictive and prescrip-
tive to better understand what may happen
LQ�WKH�IXWXUH�DQG�KRZ�WKH\�VKRXOG�UHVSRQG��'HSHQGLQJ�RQ�WKH�W\SH�RI�DQDO\VLV�WKH\�ZDQW�to conduct — real-time or historical — there
DUH�GLHUHQW�GDWD�JDWKHULQJ�DQG�DQDO\WLFV�
WRROV�WR�VXSSRUW�WKRVH�HRUWV��/HWȇV�WDNH�D�closer look at each of these examples to
better understand what they entail and how
WKH\�EHQHȴW�JRYHUQPHQW�DJHQFLHV��
3UHGLFWLYH�DQG�SUHVFULSWLYH�DQDO\WLFV��ΖWȇV�hard to say with certainty what will happen
D�PRQWK�IURP�QRZ�RU�HYHQ�D�\HDU�IURP�QRZ��But agencies can use the data they have to
EHWWHU�SUHGLFW�ZKDW�PD\�KDSSHQ�LQ�WKH�IXWXUH��7KDWȇV�WKH�UHDO�YDOXH�RI�SUHGLFWLYH�DQDO\WLFV��ΖW�helps agencies plan accordingly in a variety of
DUHDV��ZKHWKHU�LWȇV�EXGJHWLQJ��IXWXUH�FXVWRP-
er demands for services or internal human
UHVRXUFH�SODQQLQJ��%XW�NQRZLQJ�ZKDW�PD\�KDSSHQ�GRHV�DQ�DJHQF\�QR�JRRG�LI�LWȇV�XQSUH-
SDUHG�WR�UHVSRQG��7KDWȇV�ZKHUH�SUHVFULSWLYH�DQDO\WLFV�FRPHV�LQWR�SOD\��
“If agencies are able to predict the probability
of an event or action occurring, that infor-
PDWLRQ�FDQ�EH�XVHG�WR�GULYH�WKHLU�GHFLVLRQV�ȋ�+LOO�VDLG��Ȋ3UHVFULSWLYH�DQDO\WLFV�KHOSV�GHFL-VLRQ�PDNHUV�ȴJXUH�RXW�ZKDW�WR�GR�ZLWK�WKH�UHVXOWV�IURP�WKHLU�PRGHOV��7R�PDNH�GHFLVLRQV�you have to be informed, and the better your
LQIRUPDWLRQ��WKH�EHWWHU�\RXU�GHFLVLRQV�ȋ
6WUHDPLQJ�YV��EDWFK�DQDO\WLFV��6WUHDPLQJ�and batch refer to the types of analytics
WHFKQRORJLHV�WKDW�DUH�XVHG�WR�GULYH�VSHFLȴF�RXWFRPHV��)RU�H[DPSOH��VWUHDPLQJ�WHFKQROR-
gies are usually used to power real-time data
analysis that require immediate data pro-
FHVVLQJ�DQG�LQVLJKWV��6WUHDPLQJ�WHFKQRORJLHV�involve the use of sensors to rapidly collect
WKH�GDWD�WKDW�DJHQFLHV�PXVW�WKHQ�DQDO\]H��
Batch analytics, on the other hand, is used in
VFHQDULRV�WKDW�GRQȇW�UHTXLUH�LPPHGLDWH�GHFL-VLRQ�PDNLQJ��ΖQ�WKLV�LQVWDQFH��DJHQFLHV�KDYH�more time to tweak their models for more
DFFXUDWH�LQVLJKWV��
Ȋ7KHUH�LV�D�JUHDWHU�IRFXV�RQ�TXDOLW\�LI�\RXȇUH�doing batch, where as with real-time analysis
there are online algorithms making decisions
DV�WKH\ȇUH�UXQQLQJ�ȋ�+LOO�VDLG��
5HDO�WLPH�YV��IRUHQVLF�DQDO\VLV��1RZ�WKDW�ZHȇYH�GLVFXVVHG�VRPH�RI�WKH�W\SHV�RI�DQDO\WLFV�DJHQFLHV�FDQ�XVH��OHWȇV�H[SORUH�WKH�RXWFRPHV�WKH\�SURGXFH��7R�UHLWHUDWH��6WUHDP-
ing analytics supports real-time analysis, while
EDWFK�DQDO\WLFV�VXSSRUWV�IRUHQVLF�DQDO\VLV��
Both real-time and forensic analyses provide
great value to agencies, depending on their
VSHFLȴF�UHTXLUHPHQWV��+LOO�VDLG��5HDO�WLPH�analysis is helpful when decisions are based
on quickly evolving data, such as current
WUDɝF�RQ�WKH�URDG�RU�FKDQJLQJ�ZHDWKHU�FRQGLWLRQV��%XW�EDWFK�DQDO\WLFV�ZRXOG�EH�WKH�more appropriate option for driving forensic
analyses looking at historical data, such as
WUDɝF�SDWWHUQV�RYHU�WLPH�RU�ZHDWKHU�FRQGL-WLRQV�RYHU�WKH�SDVW�ZHHN�RU�VHYHUDO�PRQWKV��
ΖQ�DQ�HPHUJHQF\�UHVSRQVH�VLWXDWLRQ��ȴUVW�responders would rely on real-time analysis
to determine how far an ambulance is from
a victim, and how far the victim is from the
QHDUHVW�KRVSLWDO��$�IRUHQVLF�DQDO\VLV�RI�WKDW�same data may be used to better understand
which parts of the city receive the most 9-1-1
calls and where ambulances should be sta-
WLRQHG�WR�EHWWHU�UHVSRQG�WR�WKRVH�FDOOV��
�Ȋ:HȇUH�DOZD\V�LQWHUHVWHG�LQ�WKLQJV�WKDW�DUH�just out of reach, and some people think
WKLV�WHFKQRORJ\�LV�RXW�RI�UHDFK�ȃ�EXW�LWȇV�QRW�ȋ�+LOO�VDLG��7KDQNV�WR�GDWD�DQDO\WLFV��agencies have all of these capabilities and
PRUH�DW�WKHLU�ȴQJHUWLSV��
12 | A GOVLOOP GUIDE
Government Accountability OfficeAn interview with Seto BagdoyanDirector of Audit Services for GAO’s Forensic Audits & Investigative Service
7UDFNLQJ�GRZQ�IUDXG�LQ�WKH�JRYHUQPHQWȇV�PXOWLELOOLRQ�GROODU�0HGLFDUH�SURJUDP�LV�QRW�IRU�WKH�IDLQW�RI�KHDUW��
&RQVLGHU�WKLV��0HGLFDUH�VKHOOHG�RXW������ELOOLRQ�IRU�KHDOWKFDUH�DQG�RWKHU�UHODWHG�VHUYLFHV�LQ�ȴVFDO�������%XW�����ELOOLRQ�ȃ�RU�DERXW����percent — of that amount was paid improperly, according to estimates
IURP�WKH�&HQWHUV�IRU�0HGLFDUH��0HGLFDLG�6HUYLFHV��&06��
Ineligible providers and suppliers intent on defrauding the gov-
ernment have used all sorts of schemes to enroll in the Medicare
SURJUDP��'HVSLWH�&06ȇ�PXOWLOD\HUHG�VFUHHQLQJ�SURFHGXUHV�WR�ZHHG�RXW�IUDXG��VRPH�DSSOLFDQWV�WU\�WR�H[SORLW�D�VSHFLȴF�ZHDNQHVV�LQ�WKH�DJHQF\ȇV�YHULȴFDWLRQ�SURFHVV�E\�OLVWLQJ�TXHVWLRQDEOH�DGGUHVVHV�WKDW�DUH�QRW�ȊRSHUDWLRQDOȋ�WR�VXSSRUW�0HGLFDUH�FRYHUHG�LWHPV�DQG�VHU-YLFHV��DV�UHTXLUHG�E\�IHGHUDO�UHJXODWLRQV�
A June 2015 report�E\�WKH�*RYHUQPHQW�$FFRXQWDELOLW\�2ɝFH�IRXQG�addresses listed as legitimate practice locations that were later traced
to mailbox rental stores, a construction site, a fast-food franchise and
D�ZHOO�OLW�EXW�HPSW\�RɝFH�
Ȋ:H�IRXQG�WKDW�WKHUH�ZHUH�VRPH�VLJQLȴFDQW�LVVXHV�WKDW�WKH�DJHQF\�DQG�LWV�FRQWUDFWRUV�ZHUH�PLVVLQJ�ȋ�VDLG�6HWR�%DJGR\DQ��'LUHFWRU�RI�$XGLW�6HUYLFHV�IRU�*$2ȇV�)RUHQVLF�$XGLWV�DQG�ΖQYHVWLJDWLYH�6HUYLFH��DQG�OHDG�DXWKRU�RI�WKH�UHSRUW�
But how do these blatant cases of fraud slip under the radar? Ac-
FRUGLQJ�WR�*$2��RQH�UHDVRQ�LV�WKDW�WKH�VRIWZDUH�&06�XVHV�WR�YHULI\�DGGUHVVHV�GRHV�QRW�ȵDJ�LOOHJLWLPDWH�ORFDWLRQV�
)RU�*$2��UHYLHZLQJ�WKH�GDWD�DQG�ZHHGLQJ�RXW�IUDXGXOHQW�SURYLGHUV�DQG�suppliers, particularly those with ineligible addresses, requires months
RI�LQWHQVH�ZRUN��
ȊΖWȇV�D�YHU\�PHWKRGLFDO��RIWHQ�WHGLRXV�SURFHVV�WR�REWDLQ�WKH�GDWD��GHWHU-PLQH�LWV�UHOLDELOLW\��ȴJXUH�RXW�ZKDW�VRUW�RI�PDWFKLQJ�RU�PLQLQJ�\RXȇUH�going to perform, and then write the appropriate program using tools
VXFK�DV�6$6�DQG�ΖGHD�WR�DFWXDOO\�SHUIRUP�WKH�DQDO\VLV��JHW�WKH�LQLWLDO�results, and verify those results through due diligence before reaching
\RXU�DQDO\WLFDO�FRQFOXVLRQV�ȋ�%DJGR\DQ�VDLG��
THE CHALLENGE: Tracking Medicare Fraud
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 13
1 2 3
$V�%DJGR\DQ�PHQWLRQHG��WKHUHȇV�D�ORW�RI�SUHSDUDWLRQ�EHKLQG�WKH�VFHQHV�EHIRUH�GDWD�FDQ�EH�SURSHUO\�DQDO\]HG��
+LV�WHDP�ZRUNHG�ZLWK�VSHFLDOLVWV�LQ�*$2ȇV�$SSOLHG�5HVHDUFK�DQG�0HWK-
RGȇV�JURXS�WR�ZULWH�FRGH�WKDW�HQDEOHG�WDUJHWHG�DQDO\VLV�RI�TXHVWLRQ-
DEOH�SURYLGHU�DGGUHVVHV�VXEPLWWHG�WR�0HGLFDUH��
7KH�WHDP�DOVR�XVHG�D�FRPPHUFLDOO\�DYDLODEOH�VRIWZDUH�SDFNDJH�IURP�WKH�8�6��3RVWDO�6HUYLFH�WKDW�VWDQGDUGL]HV�DGGUHVVHV�DQG�ȵDJV�ORFDWLRQV�WKDW�DUHQȇW�OHJLWLPDWH��%DJGR\DQ�H[SODLQHG�WKDW�YDOLGDWLQJ�WKHVH�ORFD-
WLRQV�LV�D�YHU\�GHOLEHUDWH�SURFHVV��
ȊΖWȇV�QRW�OLNH�\RX�WKURZ�D�PLOOLRQ�RU�VR�0HGLFDUH�SURYLGHU�QDPHV�LQWR�WKH�DGGUHVV�GDWDEDVH�DQG�LW�DOO�FRPHV�RXW�ZLWK�UHVXOWV�ȋ�KH�VDLG��Ȋ<RX�KDYH�WR�FRGH�DQG�SURJUDP�WKH�VRIWZDUH�WR�DFFHSW�WKH�GDWDEDVH�DQG�WKHQ�ERXQFH�WKHP�R�RI�HDFK�RWKHU�WR�WKURZ�R�WKH�OHDGV�WKDW�\RX�ZLOO�WKHQ�WDNH�RXW��DQG�FRQȴUP�WKURXJK�ZKDW�ZH�FDOO�WUDFLQJ�DQG�YHULI\LQJ��ZKLFK�LV�D�YHU\�WHGLRXV�H[HUFLVH��ΖWȇV�SDUW�RI�GXH�GLOLJHQFH�WR�make sure that the emerging results are indeed results, rather than
SURJUDPPLQJ�JOLWFKHV�ȋ
ΖQ�LWV�0HGLFDUH�UHSRUW��*$2�SDUVHG�QHDUO\���PLOOLRQ�DGGUHVVHV�XVLQJ�WKH�8636�GDWDEDVH�DQG�LGHQWLȴHG�PRUH�WKDQ���������RI�WKHP�DV�
SRWHQWLDOO\�LQHOLJLEOH��8VLQJ�D�UDQGRP�VDPSOH�RI�ȵDJJHG�DGGUHVVHV��WKH�DJHQF\�HVWLPDWHG�WKDW��������GLGQȇW�PHHW�IHGHUDO�UHTXLUHPHQWV��
Ȋ7KDW�ZDV�RXU�GHHS�GLYH�ȋ�%DJGR\DQ�VDLG��*$2�DOVR�XVHG�*RRJOH�0DSV�WR�WUDFN�GRZQ�YDULRXV�ORFDWLRQV�DQG�VHQW�WHDPV�IURP�LWV�ȴHOG�RɝFHV�WR�YLVLW�TXHVWLRQDEOH�VLWHV��
“By looking at a lot of data and drilling down to the appropriate level, we
KDYH�UHDVRQDEOH�DVVXUDQFH�WKDW�ZHȇYH�H[KDXVWHG�ZKDW�ZH�FDQ�GR��DQG�ZH�FDQ�EDVH�DQ\�FRQFOXVLRQV�RQ�WKH�ZRUN�WKDW�ZHȇYH�GRQH�ȋ�KH�QRWHG��
*$2�VXFFHVVIXOO\�FRPELQHG�WKH�SRZHU�RI�GDWD�DQDO\VLV�VRIWZDUH��UHDG-
ily available Internet-based tools and on-the-ground follow-up to verify
LWV�ȴQGLQJV��ΖW�GLGQȇW�WDNH�DQ�DUP\�RI�H[SHUWV��ΖQ�IDFW��DQ�LQWHUQ�FRQGXFW-HG�PRVW�RI�WKH�*RRJOH�0DSV�DQDO\WLFV�ZRUN��
Ȋ:HȇUH�D�YHU\�VPDOO�DJHQF\��DQG�P\�WHDP��WKH�)RUHQVLF�$XGLWV�DQG�ΖQYHVWLJDWLYH�6HUYLFH��LV�HYHQ�VPDOOHU�ȋ�%DJGR\DQ�VDLG��ȊΖWȇV�WKH�VPDOOHVW�PLVVLRQ�WHDP�LQ�*$2��:HȇUH�ULJKW�DURXQG����RU�VR�SHRSOH��DQG�ZH�punch way above our weight because we have the determination to
GR�D�JRRG�MRE�ZLWK�WKH�GDWD�DQG�WKH�WRROV�ZH�KDYH�ȋ
SUCCESS FACTORS
THE SOLUTION
Ensure data is clean, usable, & reliable as it’s collected.
Institute multiple levels of review to double & triple check that the
data is secure.
Adhere to policies & regulations that govern data collection & publication.
“We’re a very small agency, and my team...we punch way above our weight because we have the determination to do a good job with the data and the tools we have.”
$ 5 5 4 , 0 0 0 , 0 0 0 , 0 0 0 paid by Medicare for healthcare & related services in FY2014
11 percent ($60 billion!) was paid improperly
23,4001 teamof GAO employees
1million addresses
addresses that did not meet federal requirements
sorted through
and found
14 | A GOVLOOP GUIDE
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 15
Accelerating Insights with Multi-Tiered Flash Storage An interview with Chris Tsilipounidakis & Jason StrawdermanManager of Product Marketing and Vice President, Federal, at Tegile
INDUSTRY SPOTLIGHT
At its core, big data analytics is about
converting data into knowledge that can be
used to make informed decisions, which
can have a variety of positive implications to
JRYHUQPHQW�RSHUDWLRQV��
)RU�LQVWDQFH��ELJ�GDWD�LV�EHLQJ�XVHG�LQ�F\EHU�defense to accelerate recognition of pat-
WHUQV�WKDW�UHSUHVHQW�WKUHDWV�WR�WKH�QHWZRUN��7KHVH�LQVLJKWV�DUH�HQDEOLQJ�RUJDQL]DWLRQV�of all types to predict and prevent potential
cyberattacks before they occur, said Jason
6WUDZGHUPDQ��9LFH�3UHVLGHQW�RI�)HGHUDO�DW�7HJLOH��D�&DOLIRUQLD�EDVHG�FRPSDQ\�WKDW�VSH-
FLDOL]HV�LQ�QH[W�JHQHUDWLRQ�PXOWL�WLHUHG�ȵDVK�VWRUDJH�V\VWHPV�
“Government healthcare organizations are
using big data to manage, analyze, visual-
ize, and extract information that is helping
SUHYHQW�DQG�ȴJKW�GLVHDVH�ȋ�6WUDZGHUPDQ�VDLG��Ȋ:KLOH�WKH�EHQHȴWV�RI�ELJ�GDWD�LQ�RXU�government are tremendous, the implemen-
tation is proving very challenging as legacy
infrastructure cannot handle the demanding
workload, and disparate systems prevent the
VKDULQJ�RI�LQIRUPDWLRQ��%LJ�GDWD�LV�IRUFLQJ�RXU�JRYHUQPHQW�WR�PRGHUQL]H�WKHLU�Ζ7�LQIUDVWUXF-ture and transform culture, policies and pro-
FHVVHV��7KHVH�FKDOOHQJHV�DUH�QRW�MXVW�OLPLWHG�to the government; many organizations in
WKH�SULYDWH�VHFWRU�IDFH�WKHVH�LVVXHV�DV�ZHOO�ȋ�6WUDZGHUPDQ�VDLG��
With this transformation comes the need
IRU�D�GLHUHQW�ZD\�RI�VWRULQJ�DQG�PDQDJLQJ�ELJ�GDWD��$JHQFLHV�PXVW�EUHDN�GRZQ�OHJDF\�storage infrastructure silos while improving
performance and increasing capacity, said
&KULV�7VLOLSRXQLGDNLV��0DQDJHU�RI�3URGXFW�0DUNHWLQJ�DW�7HJLOH��
&XVWRPHUV�DUH�PRYLQJ�DZD\�IURP�WKH�LGHD�RI�SXUSRVH�EXLOW�VWRUDJH��ZKHUH�WKH\�XVH�GLHU-HQW�YHQGRUV�WR�VWRUH�GLHUHQW�W\SHV�RI�VWUXF-WXUHG�DQG�XQVWUXFWXUHG�GDWD��7VLOLSRXQLGDNLV�H[SODLQHG��ΖQ�WKH�SDVW��RUJDQL]DWLRQV�W\SLFDOO\�used certain vendors to store mission-
critical data on more expensive, high-pow-
ered storage systems, other vendors to store
less important data on lower-cost storage,
while using more general-purpose systems
for workloads that require an even blend of
SHUIRUPDQFH��FDSDFLW\�DQG�HFRQRPLFV��
“Big data analytics is causing a massive
K\SHU�FRQVROLGDWLRQ�LQ�WKH�GDWDFHQWHU�ȋ�KH�H[SODLQHG��Ȋ+\SHU�FRQVROLGDWLRQ�PHDQV�WKDW�LI�ΖȇYH�JRW�OHVV�PRQH\��OHVV�UHVRXUFHV��EXW�more requirements to do the same thing
that I was doing before, I need to do it in a
FRQVROLGDWHG�IDVKLRQ�ȋ
7UDGLWLRQDOO\��OHJDF\�VWRUDJH�V\VWHPV�SRSX-
lated with a mix of solid-state and hard-disk
GULYHV�KDYH�EHHQ�WKH�JR�WR�IRU�DJHQFLHVȇ�VWRU-DJH�QHHGV��%XW�WR�KHOS�DJHQFLHV�EHWWHU�PHHW�the workload demands of big data analytics,
7HJLOH�LV�ZRUNLQJ�ZLWK�WKHP�WR�DGRSW�PXOWL�WLHUHG�ȵDVK�VWRUDJH��
Ȋ:LWK�D�PXOWL�WLHUHG�ȵDVK�VWRUDJH�DUFKLWHF-WXUH��Ζ�FDQ�WHOO�&;2ȇV��3URJUDP�0DQDJHUV��DQG�Ζ7�0DQDJHUV�WKDW�Ζ�FDQ�JLYH�WKHP�IDU�EHWWHU�performance than they are getting from their
legacy storage systems and at a fraction of
WKH�FRVW��ΖQ�IDFW��RXU�QHZ�+LJK�'HQVLW\�)ODVK�DUUD\V�RHU�ȵDVK�SHUIRUPDQFH�DW�WKH�VDPH�FRVW�SHU�JLJDE\WH�RI�WUDGLWLRQDO�KDUG�GULYHV�ȋ�7VLOLSRXQLGDNLV�VDLG�
7KH�7HJLOH�PXOWL�WLHUHG�ȵDVK�DUFKLWHFWXUH�XVHV�GLHUHQW�OD\HUV�RI�ȵDVK�WR�PHHW�DJHQFLHVȇ�YDU\-LQJ�QHHGV��ZKHWKHU�LWȇV�PDQDJLQJ�VWUXFWXUHG�RU�XQVWUXFWXUHG�GDWD��7KLV�DSSURDFK�SURYLGHV�D�ȵH[LEOH��H[WHQVLEOH��VFDODEOH��DRUGDEOH��DQG�HQHUJ\�HɝFLHQW�DUFKLWHFWXUH�WKDW�FDQ�DFFRP-
PRGDWH�FKDQJH�UDSLGO\��:LWK�GHQVH�ȵDVK��agencies can get better performance and
PRUH�GDWD�LQWR�D�VPDOOHU�IRRWSULQW��
Ȋ2XU�QHZ�KLJK�GHQVLW\�ȵDVK�DUUD\��ΖQWHOOL)ODVK�+'��FDQ�SDFN�XS�WR�IRXU�SHWDE\WHV�RI�UDZ�ȵDVK�LQ�D�VLQJOH���8�UDFN��DW�QHDU�WKH�FRVW�per gigabyte of traditional hard drives, and
LQFUHGLEO\�SRZHU�HɝFLHQW�ȋ�VDLG�6WUDZGHUPDQ�
In terms of data management, security is also
WRS�RI�PLQG�IRU�DJHQFLHV��7KDWȇV�ZK\�7HJLOH�XVHV�VHOI�HQFU\SWLQJ�ȵDVK�GULYHV�WKDW�DUH�)Ζ36�������YDOLGDWHG��PHDQLQJ�WKH\�PHHW�8�6��federal standards for encrypting data at rest
DQG�LQ�WUDQVLW��
Ȋ7HJLOH�FDQ�DGDSW�WR�ȴW�DQ\�VWRUDJH�HQYLURQ-
PHQW�DQG�ZRUNORDG�ȋ�6WUDZGHUPDQ�VDLG��“Most government agencies are migrating to
— or already operating — hybrid cloud archi-
tectures consisting of both private cloud and
WUDGLWLRQDO�Ζ7��:H�FDQ�LQWHURSHUDWH�LQ�WKRVH�existing environments, without them having
WR�ULS�DQG�UHSODFH�DOO�WKH�OHJDF\�WHFKQRORJ\�ȋ
When it comes to choosing cloud or tradi-
WLRQDO�Ζ7��LW�LVQȇW�DQ�DOO�RU�QRWKLQJ�GHFLVLRQ��6WUDZGHUPDQ�DGGHG��
However, one area where a model like cloud
can be less viable is when agencies want to
process data in real-time and take quick ac-
tions, as opposed to reviewing large amounts
of historical data for less time-sensitive
DQDO\VHV��7VLOLSRXQLGDNLV�QRWHG��7KDWȇV�ZK\�DJHQFLHV�PXVW�ȴUVW�HYDOXDWH�WKHLU�QHHGV�DQG�VHH�ZKHUH�GLHUHQW�WHFKQRORJLHV�FDQ�VXSSRUW�WKRVH�REMHFWLYHV��
6WUDZGHUPDQ�LV�DQ�DGYRFDWH�IRU�WKH�VR�FDOOHG�ȊODQG�DQG�H[SDQG�DSSURDFK�ȋ�ZKHUH�DJHQFLHV�FDQ�VWDUW�E\�XVLQJ�ȵDVK�VWRUDJH�IRU�WKHLU�ELJ�GDWD�QHHGV�DQG�RYHU�WLPH�H[SDQG�ȵDVK�WR�other workloads, such as virtualization, virtual
desktop infrastructure and archives and
backup, as older legacy systems are retired
DQG�UHIUHVKHG�
Ȋ1RERG\�ZDQWV�WR�UHSODFH�WKHLU�HQWLUH�infrastructure, and no one has the budget
IRU�WKDW�ȋ�6WUDZGHUPDQ�VDLG��Ȋ%XW�ELJ�GDWD�analytics, server consolidations, desktop
virtualization and mobility projects are a few
JUHDW�RSSRUWXQLWLHV�WR�LQWURGXFH�ȵDVK�LQWR�WKHLU�HQYLURQPHQW�ȋ
16 | A GOVLOOP GUIDE
General Services AdministrationAn interview with Paul TsagaroulisDirector of Human Capital Analytics at GSA
'DWD�FDQ�WHOO�XV�D�ORW�DERXW�WKH�JRYHUQPHQW�ZRUNIRUFH�ȃ�ZKR�ZDV�KLUHG��ZKR�OHIW��ZK\�WKH\�PD\�KDYH�OHIW��DQG�ZKR�HOVH�LV�OLNHO\�WR�OHDYH��
%XW�LWȇV�KDUG�WR�JOHDQ�WKRVH�LQVLJKWV�ZLWK�GDWD�GUDZQ�IURP�GLVSDUDWH�V\VWHPV��LQ�GLHUHQW�IRUPDWV��$QG�WKRVH�ZKR�QHHG�WKH�GDWD�WKH�PRVW�FDQȇW�HDVLO\�DFFHVV�LW��VDLG�3DXO�7VDJDURXOLV��'LUHFWRU�RI�+XPDQ�&DSLWDO�$QDO\WLFV�DW�*6$�
$OWKRXJK�*6$�KDV�D�VLQJOH�KXPDQ�UHVRXUFHV�V\VWHP��LW�FRPSULVHV�PXO-WLSOH�VXEV\VWHPV�WKDW�GRQȇW�WDON�WR�HDFK�RWKHU��7KLV�RQO\�H[DFHUEDWHV�WKH�FKDOOHQJH�RI�FRPELQLQJ�GLVSDUDWH�GDWD�VHWV�WR�WHOO�D�FRPSOHWH�VWRU\��7R�VWD\�FXUUHQW��*6$�KDV�KDG�WR�EXLOG�VHSDUDWH�SURFHVVHV�WR�FDSWXUH�QHZ�+5�GDWD�VHWV��VXFK�DV�WHOHZRUN�GDWD�
“A big challenge for us is data management and information manage-
PHQW�ȋ�7VDJDURXOLV�H[SODLQHG��Ȋ:H�VSHQG�D�ORW�RI�RXU�WLPH�IRFXVLQJ�RQ�JHWWLQJ�DFFHVV�WR�WKH�GDWD��PDNLQJ�VXUH�LWȇV�YDOLG�DQG�UHOLDEOH��DQG�WKDW�leaves very little time for us to analyze the data and look at trends and
patterns, maybe do some sophisticated analytics, and some predictive
DQG�SUHVFULSWLYH�DQDO\VHV�ȋ
%XW�WKDWȇV�FKDQJLQJ��WKDQNV�WR�WKH�ZRUN�7VDJDURXOLV�DQG�KLV�WHDP�RYHUVHHV�
3ULRU�WR�*6$ȇV������FRQVROLGDWLRQ�RI�LWV�+5�IXQFWLRQV��LQHɝFLHQF\�ZDV�D�ELJ�SUREOHP��7KHUH�ZHUH�RɝFHV�ZLWKLQ�HDFK�RI�*6$ȇV�EXVLQHVV�OLQHV��and they all performed the same data analytics services, multiple
WLPHV��IRU�GLHUHQW�DJHQF\�FXVWRPHUV��
Ȋ*6$�FRQVROLGDWHG��DQG�QRZ�ZHȇYH�EHHQ�DEOH�WR�WDNH�D�WRS�GRZQ�DS-
SURDFK�LQ�WKLV�SURFHVV�ȋ�7VDJDURXOLV�VDLG��Ȋ6R�ZH�GR�LW�RQFH��ZH�GHOLYHU�LW�RQFH��EXW�ZH�FDQ�DGMXVW�LW�IRU�DQ\�EXVLQHVV�OLQHȇV�VSHFLȴF�QHHGV��JHRJUDSKLF�QHHGV�RU�D�FRPELQDWLRQ�RI�WKRVH��:H�ZHUH�FUHDWHG�DV�D�GL-vision to support this for the agency, to alleviate the need for business
OLQHV�WR�GR�WKLV�WKHPVHOYHV�ȋ
THE CHALLENGE: Improving Human Capital Management
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 17
1 2 3
ΖQ�������*6$�ODXQFKHG�D�GLYLVLRQ�GHGLFDWHG�WR�GHYHORSLQJ�ZRUNIRUFH�DQDO\WLFV�GHOLYHUDEOHV��7KH�ȴUVW�RI�WKRVH�GHOLYHUDEOHV�ZDV�DQ�LQWHUDFWLYH�GDVKERDUG�WKDW�YLVXDOL]HG�DQVZHUV�WR�FRPPRQ�+5�TXHVWLRQV�DERXW�new hires and separations by grade, onboarding by major job series,
DQG�PRUH�
“We dramatically reduced ad hoc reporting and responding to re-
TXHVWV�DV�WKH\�FDPH�LQ�ȋ�7VDJDURXOLV�VDLG��
7KH�GDVKERDUG�GDWD�KDV�DOUHDG\�LQȵXHQFHG�GHFLVLRQ�PDNLQJ�DW�WKH�DGPLQLVWUDWRU�OHYHO��$IWHU�FXWWLQJ�WKH�QXPEHU�RI�DQQXDO�H[WHUQDO�KLUHV�IURP�DERXW�����LQ�ȴVFDO�\HDUV������DQG������WR�DERXW�����LQ�������*6$�ZDV�LQ�D�ELQG��
7KH�DJHQF\�ȴOOHG�IHZHU�SRVLWLRQV�EXW�DW�KLJKHU�*HQHUDO�6FKHGXOH�JUDGH�OHYHOV��ZKLFK�PHDQW�KLJKHU�FRVWV��7KH�GDVKERDUG�FRQȴUPHG�ZKDW�PDQ\�VXVSHFWHG��%\�KLULQJ�SHRSOH�LQ�PRUH�VHQLRU�JUDGHV��WKH�agency was increasing its costs and diminishing its talent development
SLSHOLQH�RI�\RXQJHU�KLUHV�FRPLQJ�LQ�DW�ORZHU�JUDGHV��7R�DGGUHVV�WKH�LPEDODQFH��*6$�LPSOHPHQWHG�DQ�HQWU\�OHYHO�KLULQJ�SURJUDP��ΖQ�ȴVFDO�������WKH�DJHQF\�WULSOHG�WKH�QXPEHU�RI�HPSOR\HHV�LW�KLUHG�DW�WKH�*6���OHYHO��FRPSDUHG�ZLWK�WKH�SUHYLRXV�\HDU��
Ȋ7KH�GDVKERDUG�ZDV�XVHIXO�LQ�VWDUWLQJ�WKH�FRQYHUVDWLRQ�ȋ�7VDJDURXOLV�VDLG��
(DUO\�RQ��KLV�WHDP�H[DPLQHG�WUHQGV�DFURVV�*6$ȇV�ZRUNIRUFH�GDWD��5HYLHZLQJ�WKRVH�WUHQGV�OHG�WKH�WHDP�WR�DVN�VSHFLȴF�TXHVWLRQV�DERXW�WXUQRYHU�UDWHV��ZKDW�IDFWRUV�FRUUHODWH�WR�WXUQRYHU��DQG�LI�*6$�FRXOG�SUHGLFW�WXUQRYHU��7KLV�OHG�LQ�WXUQ�WR�WKH�GHYHORSPHQW�RI�D�WXUQRYHU�PRGHO�IRU�SUHGLFWLQJ�IXWXUH�HPSOR\HH�ORVVHV��
Ȋ2QH�RI�WKH�ȴQGLQJV�WKDW�ZH�FDPH�DFURVV�ZDV�WKDW�SHRSOH�LQ�WKH�FHQWUDO�RɝFH�LQ�WKH�'�&��DUHD�DUH�PRUH�OLNHO\�WR�OHDYH�IRU�DQRWKHU�DJHQF\��ZKHUHDV�LQGLYLGXDOV�LQ�WKH�UHJLRQV�DUH�PRUH�OLNHO\�WR�UHWLUH�ȋ�7VDJDURXOLV�H[SODLQHG��Ȋ6R�ZH�FDQ�DFFRXQW�IRU�VRPH�GLHUHQFHV�LQ�WXUQRYHU�EDVHG�RQ�JHRJUDSKLF�YDULDEOHV��7KHQ��ORRNLQJ�IXUWKHU�LQWR�WXUQRYHU�GDWD��ZH�FDQ�SUHGLFW�ZLWK�VRPH�FRQȴGHQFH�WKH�QXPEHU�RI�people we would expect to have retire or voluntarily separate from
the agency, and we could calculate the cost of the attrition savings
RYHU�WKH�FRXUVH�RI�WKH�\HDU�ȋ
)URP�WKHUH��WKH�DJHQF\�FDQ�XVH�WKRVH�HVWLPDWHV�WR�GR�VRPH�UHFUXLWLQJ�DKHDG�RI�WLPH��LQ�DQWLFLSDWLRQ�RI�WXUQRYHU��
2YHUDOO��WKH�GLYLVLRQ�KDV�LPSURYHG�KRZ�*6$�PDQDJHV�LWV�+5�GDWD�VHWV�by automating reporting of turnover, onboarding, and performance
GDWD��%XLOGLQJ�WKDW�IRXQGDWLRQ�KDV�PDGH�LW�HDVLHU�IRU�WKH�GLYLVLRQ�WR�employ prescriptive analytics and help managers develop action plans
IRU�VWDɝQJ��7VDJDURXOLVȇ�WHDP�LV�LQ�WKH�SURFHVV�RI�UROOLQJ�RXW�DQ�LQWHUQDO�+5�GDWD�ZDUHKRXVH�WR�FRQVROLGDWH�GDWD�DQG�LPSURYH�GDWD�VKDULQJ��
SUCCESS FACTORS
THE SOLUTION
Tackle data management issues first.Use visualization techniques to explain
trends & other findings in the data.
Let workforce data dictate new HR solutions or programs that maximize
employee potential.
“Looking further into turnover data, we can predict with some confidence the number of people we would expect to have retire or voluntarily separate from the agency.”
18 | A GOVLOOP GUIDE
Beyond Big Data to Data-Driven Insights
More effective government sits at the intersection of insight and action.
When you can collect, unify, and analyze all the data that surrounds your department, government is empowered to uncover key insights that matter most. It could be an insight that changes the way citizens view your agency, or how your department sees their role. Learn how big data and analytics for smart government can help you know more so you can do more in this informative white paper.
Whitepaper: Analytics in Service of Smart Government
What would you do if you knew? is a trademark, and Teradata and the Teradata logo are registered trademarks of Teradata Corporation and/or its affiliates in the U.S. and worldwide.
What would you do if you knew?™
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 19
The Future of Data Analytics in Government ITAn interview with Alan FordDirector of Presales Consulting, Teradata Government Systems
INDUSTRY SPOTLIGHT
:KDW�GRHV������ORRN�OLNH�IRU�JRYHUQPHQW�Ζ7"�We may be in for a few surprises, but one
WKLQJ�LV�FOHDU��GDWD�DQDO\WLFV�ZLOO�SOD\�D�NH\�UROH��7KHUH�DUH�VHYHUDO�DUHDV�ZKHUH�GDWD�DQG�analytics will drive more value in government,
including data transparency, tax fraud, and
WUDQVSRUWDWLRQ�ORJLVWLFV��
'DWD�DQDO\WLFV�ZLOO�UHTXLUH�D�FRQWLQXHG�HRUW�LQ�DQDO\]LQJ�LQFUHDVLQJO\�ODUJHU�GDWDVHWV��$QG�these larger datasets are coming at govern-
ment agencies at greater speeds than ever
EHIRUH��ZLWK�ZLGHU�YDULHW\�DQG�FRPSOH[LW\��
7KLV�LV�ZKHUH�FKDOOHQJHV�EHJLQ�WR�RFFXU��:LWK�more data to analyze, government agen-
cies can easily be overwhelmed as they try
to maintain system performance and data
integrity while protecting sensitive informa-
WLRQ��$GGLWLRQDOO\��ZLWK�PRVW�DJHQFLHV�KDYLQJ�OHJDF\�V\VWHPV�LQ�SODFH��LW�FDQ�EH�GLɝFXOW�WR�LQWHJUDWH�QHZ�V\VWHPV�DQG�GLHUHQW�YHQGRUV�LQWR�WKHLU�H[LVWLQJ�DQDO\WLF�HFRV\VWHP�
7KHUH�DUH�ZD\V�WR�FRPEDW�WKHVH�FKDOOHQJHV��ΖQ�DQ�LQWHUYLHZ�ZLWK�*RY/RRS��$ODQ�)RUG��'LUHFWRU�RI�3UHVDOHV�&RQVXOWLQJ�IRU�7HUDGDWD�*RYHUQPHQW�6\VWHPV��VKDUHG�KRZ�DJHQFLHV�can leverage data analytics, including those in
the cloud, to generate actionable information
and enable better data sharing within and
DFURVV�DJHQFLHV�
7HUDGDWD�SURYLGHV�HQG�WR�HQG�VROXWLRQV�DQG�services like data warehousing to enable
JRYHUQPHQW�DJHQFLHV�WR�FUHDWH�XQLȴHG�GDWD�HFRV\VWHPV��7KH�7HUDGDWD�8QLȴHG�'DWD�$UFKLWHFWXUH��IRU�H[DPSOH��LQWHJUDWHV�GLHUHQW�platforms into a comprehensive analytics
solution that enables fast, deep, and powerful
GDWD�PDQDJHPHQW��VWRUDJH��DQG�H[SORUDWLRQ��With this architecture and its inherent analytic
capability, agencies can complement and
augment their legacy systems rather than
DEDQGRQLQJ�WKHP�
7KHVH�WHFKQRORJLHV�ZLOO�KHOS�DOO�DUHDV�RI�JRYHUQPHQW��)RUG�FLWHG�KHDOWKFDUH�IUDXG�DV�D�PDMRU�GDWD�DQDO\WLFV�SUREOHP�IRU�JRYHUQPHQW��Healthcare comprises some of the largest
annual expenditures in both the private and
SXEOLF�VHFWRU��0HGLFDUH�DQG�0HGLFDLG��LQ�SDUWLFXODU��UHFHLYH�D�VLJQLȴFDQW�DPRXQW�RI�JRY-ernment funding, which makes it an attractive
WDUJHW�IRU�IUDXGXOHQW�DFWLYLW\��
$FFRUGLQJ�WR�)RUG��KRZHYHU��WKHUH�LV�KRSH�with increased data digitization and analytics
DXWRPDWLRQ��Ȋ'LJLWL]DWLRQ�KDV�FUHDWHG�D�QHZ�ZD\�RI�DXWRPDWLQJ�IUDXG�GHWHFWLRQ�ȋ�)RUG�VDLG��“By combining traditional claims data with
other separate but publicly available data –
like bankruptcies, liens, judgments, or even
repeated address changes – into an integrat-
HG�GDWD�PRGHO�RU�D�XQLȴHG�GDWD�DUFKLWHFWXUH��we can run advanced analytics that make
fraud easier to track or even proactively pre-
GLFW�WKH�OLNHOLKRRG�RI�LWV�RFFXUUHQFH�ȋ�
7HUDGDWDȇV�$VWHU�'LVFRYHU\�3ODWIRUP�LV�RQH�tool that helps run analytics that could only
EH�GRQH�PDQXDOO\�LQ�WKH�SDVW��Ȋ:LWK�WKLV�WRRO��we can run older transcribed handwritten
records or typed reports through complex
text analytics capabilities and sentiment
analysis in an automated fashion dramatically
GHFUHDVLQJ�WKH�WLPH�WR�UHVXOWV�ȋ�)RUG�VDLG��“And with less expensive storage accessible
by sophisticated techniques, all within the
XQLȴHG�GDWD�DUFKLWHFWXUH��ZHȇOO�EH�DEOH�WR�analyze all the data rather than just represen-
WDWLYH�VDPSOHV�RI�LW�ȋ
:KHQ�ORRNLQJ�WR�WKH�IXWXUH��)RUG�SUHGLFWV�data analytics and cloud will make a strong
SDLULQJ�IRU�JRYHUQPHQW��*RYHUQPHQWV�ZLOO�EHQHȴW�IURP�WKH�UHGXFHG�FRVWV�HQDEOHG�E\�such technologies, so agencies can do more
ZLWK�OHVV��
Ȋ7KH�ZKROH�PRYHPHQW�LQWR�WKH�FORXG�HQYLURQ-
ment is a huge cost saver, because no longer
do organizations have to set up and main-
WDLQ�WKHLU�RZQ�LQIUDVWUXFWXUH�ȋ�)RUG�VDLG��ȊΖWȇV�WDNHQ�FDUH�RI�E\�WKH�FORXG�SURYLGHU��ΖQ�WKLV�way, agencies convert acquisition costs into
RSHUDWLQJ�FRVWV�ȋ
7KLV�ELJ�OHDS�WR�FORXG�DQG�GDWD�DQDO\WLFV��KRZ-
ever, may seem improbable for government
DJHQFLHV�ZLWK�OHJDF\�V\VWHPV��:KHQ�DJHQFLHV�have spent a tremendous amount of money
acquiring systems over time, they are reluc-
WDQW�WR�UHSODFH�WKHVH�IDPLOLDU�SODWIRUPV��
7KDWȇV�ZKHUH�)RUG�VD\V�ȊWKH�+RO\�*UDLO�RI�DQDO\WLFVȋ�Ȃ�WKH�XQLȴHG�GDWD�DUFKLWHFWXUH�Ȃ�FDQ�FRPH�LQ�KDQG\��Ȋ7KLV�LV�DQ�HFRV\VWHP�which allows for the harnessing of all of an
DJHQF\ȇV�GDWD�ȋ�KH�VDLG��Ȋ:H�DOVR�UHIHU�WR�LW�DV�the logical data warehouse because it brings
the right type of analytical processing to the
XQGHUO\LQJ�GDWDVHWV��)XUWKHU��LW�DOORZV�XVHUV�to combine and analyze data from multiple
GLVSDUDWH�SODWIRUPV�DOO�LQ�D�VLQJOH�TXHU\�ȋ
A logical data warehouse can subsume older
platforms so agencies that choose not to trans-
form and migrate data into an integrated data
PRGHO�PD\�VWLOO�OHYHUDJH�WKHLU�OHJDF\�V\VWHPV�
:LWK�D�XQLȴHG�GDWD�DUFKLWHFWXUH�WKDW�FDQ�accompany legacy systems, complemented
by cloud technologies, government agencies
can better manage the vast amounts of data
DFFHOHUDWLQJ�WRZDUG�WKHP��7KH�IXWXUH�RI�GDWD�DQDO\WLFV�GRHV�QRW�KDYH�WR�EH�GDXQWLQJ��:LWK�D�XQLȴHG�GDWD�DUFKLWHFWXUH��LQVLGH�RU�RXWVLGH�of the cloud, government can be ready for
advanced analytical capabilities and harness
ODUJH�LQȵX[HV�RI�FRPSOH[�GDWD�WR�IXUWKHU�DJHQF\�LQLWLDWLYHV��
20 | A GOVLOOP GUIDE
Defense Information Systems AgencyAn interview with Dan Bart & Christopher PaczkowskiChief of Cyber Situational Awareness Systems; Cyber Situational Awareness & Analytics Division Chief
ΖPDJLQH�WKDW�\RXȇUH�DQ�DQDO\VW�DW�WKH�'HIHQVH�ΖQIRUPDWLRQ�6\VWHPV�$JHQF\��<RXU�MRE�LV�WR�GHWHFW�DQG�LQYHVWLJDWH�DQRPDORXV�QHWZRUN�WUDɝF�that may be nefarious, including malicious insiders and nation states
VHHNLQJ�WR�VWHDO�FODVVLȴHG�GDWD�DQG�ZUHDN�KDYRF�RQ�GHIHQVH�V\VWHPV��
.HHS�LQ�PLQG�WKDW�LQ�ȴVFDO������DORQH��IHGHUDO�DJHQFLHV�UHSRUWHG�QHDUO\��������VHFXULW\�LQFLGHQWV��%XW�WKDW�QXPEHU�GRHVQȇW�LQFOXGH�WKH�billions of unsuccessful hack attempts that agencies had to thwart
HDFK�PRQWK��
$V�LI�DQ�DQDO\VWȇV�ZRUN�ZHUHQȇW�KDUG�HQRXJK��LPDJLQH�KDYLQJ�WR�PRQLWRU�P\ULDG�V\VWHPV��ORRNLQJ�IRU�UHG�ȵDJV�WKDW�PD\�RU�PD\�QRW�EH�WKH�ZRUN�RI�EDG�DFWRUV��1RW�WKDW�ORQJ�DJR��'Ζ6$�DQDO\VWV�XVHG�D�PDQXDO�SURFHVV�to see the anomalous behavior and connections between various
V\VWHP�DOHUWV�
ȊΖQ�D�ORW�RI�FDVHV��LWȇV�HLWKHU�PLVVHG�RU�WKH\�JHW�WLSV�DQG�FXHV�IURP�VRPHWKLQJ�HOVH�ȋ�DOHUWLQJ�WKHP�WR�GLJ�GHHSHU��VDLG�'DQ�%DUW��&KLHI�RI�&\EHU�6LWXDWLRQDO�$ZDUHQHVV�6\VWHPV��Ȋ)RU�H[DPSOH��LI�D�ȴUHZDOO�ORJ�alerted an analyst to anomalous activities, he or she could then open
an investigation and look at other sensors to try to draw connections
IURP�WKH�GDWD�ȋ
7KDW�ZDV�WKH�ROG�ZD\�RI�GRLQJ�EXVLQHVV��
$ERXW����PRQWKV�DJR��'Ζ6$�GHSOR\HG�D�QHZ�DQRPDO\�GHWHFWLRQ�V\VWHP�WKDW�LQ�VRPH�UHVSHFWV�ZLOO�SURYLGH�WKRVH�FRQQHFWLRQV�IRU�DQDO\VWV��5DWKHU�WKDQ�VHQGLQJ�DOHUWV�WKDW�VRPHWKLQJ�PD\�QRW�EH�ULJKW��WKLV�QHZ�V\VWHP�ZLOO�OHW�DQDO\VWV�NQRZ�ZKHQ�VRPHWKLQJ�LV�QRW�ULJKW��%DUW�VDLG��
&RQVLGHU�KRZ�PDQ\�HQG�XVHUV�OLNHO\�ORJ�RQ�WR�WKHLU�FRPSXWHUV�RXWVLGH�RI�QRUPDO�EXVLQHVV�KRXUV��:LWK�WKH�ULVH�RI�WHOHZRUN�DQG�ȵH[LEOH�VFKHG-
XOHV��LWȇV�KDUG�WR�VD\�ZKDW�D�ȊQRUPDOȋ�ZRUNGD\�ORRNV�OLNH�IRU�DQ\�JLYHQ�HPSOR\HH�
But analysts have to distinguish between anomalous activity that is
QRUPDO�DQG�DEQRUPDO��
THE CHALLENGE: Empowering Cybersecurity Analysts
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 21
1 2 3
:KDWȇV�IDU�PRUH�YDOXDEOH�IRU�DQDO\VWV�LV�WKH�DELOLW\�WR�HDVLO\�FRUUHODWH�multiple data sets, such as what time users are logging on to their
computers, how that compares with their normal computer use, and
ZKDW�WKH\�ZHUH�GRZQORDGLQJ�RU�XSORDGLQJ��VDLG�3DF]NRZVNL�
Ȋ<RX�FDQ�VWDUW�WR�DGG�$��%��DQG�&�WRJHWKHU�WR�VD\��Ȇ$OULJKW��WKLV�LQGLFD-
WRU�LV�D�ORW�KLJKHU�SULRULW\�WKDQ�WKH�UHVW�RI�WKHP�ȋ�3DF]NRZVNL�DGGHG��Ȋ6R�QRZ�\RXȇUH�JLYLQJ�WKH�DQDO\VW�DQ�RUGHUHG�OLVW�RI�SULRULWLHV�WR�ORRN�LQWR�ȋ�LQVWHDG�RI�IROORZLQJ�XS�RQ�HYHU\�XVHU�ZKR�ORJJHG�LQ�DIWHU�UHJXODU�EXVLQHVV�KRXUV��
$QDO\VWV�FDQ�ZRUN�PRUH�HɝFLHQWO\�EHFDXVH�WKH\�DUHQȇW�ȵRRGHG�ZLWK�D�VHD�RI�DOHUWV�WKDW�UHDOLVWLFDOO\�FDQȇW�DOO�EH�DGGUHVVHG�DW�WKH�VDPH�WLPH��7KLV�V\VWHP�KDV�KHOSHG�'Ζ6$�SULRULWL]H�DQRPDOLHV�DQG�XQGHUVWDQG�ZKDW�GDWD�LV�OHDYLQJ�WKH�QHWZRUN�LQ�DQ�XQDXWKRUL]HG�ZD\��%DUW�VDLG��
7KH�JHQHVLV�RI�WKH�SURJUDP�KDV�VRPH�WLHV�WR�WKH�GHYDVWDWLQJ�OHDNV�E\�IRUPHU�JRYHUQPHQW�FRQWUDFWRU�(GZDUG�6QRZGHQ�DERXW�WKH�1DWLRQDO�6HFXULW\�$JHQF\ȇV�FRQWURYHUVLDO�GDWD�FROOHFWLRQ�SURFHVVHV�
%XW�LW�LVQȇW�MXVW�LQVLGHUV�WKDW�WKH�V\VWHP�FDQ�GHWHFW��7KH�VDPH�DQRPDO\�detection can be used to identify a nation state worming its way into
WKH�QHWZRUN��RU�HYHQ�D�PDFKLQH�PLVFRQȴJXUDWLRQ��
Ȋ6R�LI�\RX�KDYH�D�QXPEHU�RI�GLHUHQW�ORJ�DOHUWV�RU�GDWD�VHWV�WKDW�DUH�
ȵDJJHG��WKHQ�\RXȇUH�DEOH�WR�FRUUHODWH�WKRVH�DQG�VWDUW�DQVZHULQJ�KHX-
ULVWLF�TXHVWLRQV�ȋ�RU�QRWLQJ�SDWWHUQV�LQ�EHKDYLRU��%DUW�VDLG��
)RU�H[DPSOH��LI�DQ�DQDO\VW�VHHV�WKH�VDPH�SDWWHUQ�ZLWK�D�XVHU�ORJJLQJ�LQ�at an odd time or from an odd place, the system provides an analytic
WLS�RU�DOHUW�WKDW�ZLOO�XOWLPDWHO\�OHDG�WR�D�FRXQWHUPHDVXUH��+XPDQV�GHFLGH�ZKDW�IROORZ�XS�PHDVXUHV�WR�WDNH�ȃ�QRW�WKH�V\VWHP�
ȊΖW�GRHVQȇW�VROYH�LW�ȋ�%DUW�VDLG��ȊΖW�GRHVQȇW�FDWFK�WKH�JX\��EXW�LW�GRHV�WLS�DQG�FXH�WR�VD\��Ȇ+H\��WKLV�VKRXOG�EH�ORRNHG�DW�ȇȋ�
6LQFH�LWV�UROORXW�DERXW����PRQWKV�DJR��WKH�V\VWHP�FRQWLQXHV�WR�HYROYH��7KH�DJHQF\�LV�ZRUNLQJ�WR�ȴQH�WXQH�WKH�DOJRULWKP�WKDW�GHȴQHV�DQG�SULRULWL]HV�WKUHDW�IDFWRUV��7KLQN�EDFN�WR�WKH�HDUOLHU�H[DPSOH�RI�WKH�LQGLYLGXDO�ZKR�UHSHDWHGO\�ORJV�LQ�DIWHU�KRXUV��DW�GLHUHQW�WLPHV��:KLOH�LW�PD\�UDLVH�H\HEURZV��LWȇV�QRW�XQFRPPRQ�RU�SDUWLFXODUO\�VXVSLFLRXV�
ΖI�DQDO\VWV�FRPELQH�WKDW�DOJRULWKP�ZLWK�GLHUHQW�SULRULWLHV�VXFK�DV�GDWD�H[ȴOWUDWLRQ��ZKHUH�WKH�SHUVRQ�ORJJHG�LQ��ZKDW�GHYLFHV�WKH\�XVHG�WR�connect to the network, and other behavior, then they can combine
WKDW�LQIRUPDWLRQ�WR�JHW�D�SULRULWL]HG�UDQN�RI�DQRPDORXV�EHKDYLRU�
Ȋ$OHUWV�XVHG�WR�EH�PRUH�FRPSDUWPHQWDOL]HG��EDVHG�RQ�VSHFLȴF�EH-
KDYLRUV�ȋ�3DF]NRZVNL�VDLG��%XW�QRZ�WKH�IRFXV�LV�RQ�W\LQJ�WKDW�SLFWXUH�WRJHWKHU�WR�JLYH�DQDO\VWV�D�PRUH�FRPSOHWH�WKUHDW�SURȴOH��
SUCCESS FACTORS
THE SOLUTION
Let the use case dictate what sensors are needed and what data
should be collected.
Communicate with other people in the department, and get a feel for what
systems and capabilities they’re using.
Develop analytics in a common way, and work toward a common goal across the department, to create opportunities
to share analytics.
“If you have a number of different log alerts or data sets that are flagged, then you’re able to correlate those and start answering heuristic questions.”
LOGIN AFTER HOURS OUTSIDE U.S. FILES DOWNLOADED
SYSTEM DETECTS
ANOMALIES
SENDS PRIORITIZED
ALERTS TO ANALYSTS
ANALYSTSVIEW ALERTS& RESPOND
LOGIN AFTER HOURS
LOGIN AFTER HOURSOUTSIDE U.S.
3
2
1
22 | A GOVLOOP GUIDE
When government runs on flash
What was complex, now is simple
What filled data center racks, now fits in a box
What took days, now takes hours
What became outdated, now is evergreen
© 2015 Pure Storage, Inc. All rights reserved. Pure Storage, Pure1, and the P Logo are trademarks of Pure Storage, Inc. All other trademarks are the property of their respective owners.
Pure Storage (PSTG) is the #1 recommended technology vendor (based on Satmetrix 2015 NPS Global B2B Benchmarks). To learn more about how Pure Storage can help your agency accelerate what’s possible, visit www.purestorage.com/government
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 23
In the World of Data Analytics, Speed Is EverythingAn interview with Vaughn StewartVice President, Enterprise Architect, Pure Storage
INDUSTRY SPOTLIGHT
7KH�VSHHG�DW�ZKLFK�JRYHUQPHQW�DJHQFLHV�can access, analyze and disseminate infor-
mation impacts more than cost savings and
LQWHUQDO�HɝFLHQFLHV��
)RU�RɝFHUV�LQ�*DUODQG��7H[DV��LWȇV�DERXW�WKH�ability to quickly run queries on individuals
GXULQJ�URXWLQH�WUDɝF�VWRSV�DQG�GHWHUPLQH�what actions to take based on that informa-
WLRQ��ΖQ�PDQ\�VWDWHV��JRYHUQPHQW�RɝFLDOV�must be able to quickly detect and assess
deteriorating weather conditions and take
necessary precautions, such as closing roads
RU�WHPSRUDULO\�UHGXFLQJ�GULYLQJ�VSHHGV�
But how can agencies support the growing
demands for faster services when budgets
DUH�OLPLWHG"�)RU�D�JURZLQJ�QXPEHU�RI�DJHQ-
cies, the answer to powering these types
RI�RSHUDWLRQV�KDV�FRPH�LQ�WKH�IRUP�RI�ȵDVK�VWRUDJH��ZKLFK�FDQ�QRZ�EH�D�PRUH�FRVW�HHF-WLYH�DQG�HɝFLHQW�DOWHUQDWLYH�WR�WKH�OHJDF\�GLVN�VWRUDJH�WUDGLWLRQDOO\�XVHG�E\�DJHQFLHV��
Ȋ7KH�SRZHU�RI�ȵDVK�UHDOO\�DOORZV�DJHQFLHV�WR�DGYDQFH�WKHLU�DELOLW\�WR�JDWKHU�LQIRUPDWLRQ�ȋ�VDLG�9DXJKQ�6WHZDUW��93��(QWHUSULVH�$UFKLWHFW�DW�3XUH�6WRUDJH��ȊΖWȇV�IRFXVHG�DURXQG�FROOHFW-ing information faster to they can respond
PRUH�TXLFNO\�WR�GLHUHQW�VFHQDULRV��VXFK�DV�emergency response, cybersecurity or in the
ȴQDQFLDO�VHFWRU�WR�GHWHFW�DQG�SUHYHQW�IUDXG�ȋ�
7KH�DELOLW\�WR�SURFHVV�PRUH�GDWD�IDVWHU�DOORZV�agencies to run more queries, gain greater
insight, and have a deeper knowledge of their
JURZLQJ�GDWD��6WHZDUW�H[SODLQHG��
Much of that data, however, is stored in
YDULRXV�IRUPDWV�DFURVV�GLHUHQW�V\VWHPV��ZKLFK�FDQ�SXW�D�VWUDLQ�RQ�GDWD�VKDULQJ��
Although there are many technologies that
can provide a central repository for data, the
speed at which agencies can process that
data is another factor agencies must consid-
HU��7KDWȇV�ZKHUH�ȵDVK�VWRUDJH�KDV�FRPH�LQ�as the game changer for agencies that are
RSHQ�WR�FKDQJH�
Ȋ0DQ\�DJHQFLHV�GRQȇW�ORRN�DW�ȵDVK�XQWLO�WKH\�KDYH�D�NQRZQ�LVVXH��DQG�WKH\ȇYH�H[KDXVWHG�DOO�RWKHU�WHFKQRORJLHV�ȋ�6WHZDUW�VDLG��Ȋ8VXDOO\�WKDWȇV�EHFDXVH�ȵDVK�KDV�KLVWRULFDOO\�EHHQ�FRVW�SURKLELWLYH��%XW�WKH�UHDOLW\�LV�WKDW�ZH�DUH�LQ�DQ�DJH�ZKHUH�ȵDVK�LV�QRW�MXVW�DRUGDEOH��but in many circumstances, it costs less than
VSLQQLQJ�GLVN�VWRUDJH�ȋ�
:LWK�WKLV�LQ�PLQG��6WHZDUW�VDLG�VRPH�DJHQFLHV�DUH�VWDUWLQJ�WR�FRQVLGHU�ȵDVK�HDUOLHU�RQ��UDWK-
er than as a last resort to resolve their biggest
SDLQ�SRLQWV��)RUZDUG�WKLQNLQJ�FXVWRPHUV�XQGHUVWDQG�WKDW�PRYLQJ�WR�ȵDVK�FDQ�KHOS�them speed their work because the latency
issues often associated with disk storage
ZRQȇW�SODJXH�WKHP��
Ȋ)RU�HYHU\�PRQWK�WKDW�\RX�UXQ�RQ�ȵDVK��\RX�FDQ�FRPSOHWH�D�\HDUȇV�ZRUWK�RI�DQDO\WLFV�WKDW�DUH�FRPSOHWHG�RQ�GLVN�ȋ�6WHZDUW�VDLG�
7KHUHȇV�RQH�FDVH�LQ�SDUWLFXODU�6WHZDUW�QRWHG�WKDW�LQYROYHG�GDWD�VFLHQWLVWV�ORRNLQJ�IRU�IUDXG��$W�ȴUVW��WKH\�FRXOG�RQO\�RSHUDWH�IRXU�DG�KRF�TXHULHV�DW�D�WLPH�XVLQJ�GLVN�DUUD\V��%XW�ZKHQ�WKH\�PRYHG�WR�ȵDVK�VWRUDJH��WKDW�QXPEHU�MXPSHG�IURP�IRXU�WR�����TXHULHV��
Ȋ'DWD�VWRUDJH�LV�OLNH�D�IRXQGDWLRQDO�OD\HU�XQGHUQHDWK�DOO�RI�WKHVH�V\VWHPV�ȋ�6WHZDUW�VDLG��Ȋ7KLQN�DERXW�D�EXLOGLQJ��ΖI�\RXȇYH�JRW�D�poor foundation, you can only build a building
WR�D�FHUWDLQ�KHLJKW��VL]H�DQG�VFDOH��ΖI�\RX�FDQ�UHSODFH�D�EXLOGLQJȇV�IRXQGDWLRQ�ZLWK�D�QHZ�generation of technology or architecture
technology, you can then allow that building
WR�VFDOH�KLJKHU��'DWD�VWRUDJH�LV�YHU\�VLPLODU�ȋ
If agencies switch the infrastructure from
VSLQQLQJ�GLVN�WR�ȵDVK�PHGLD�RU�VROLG�VWDWH�storage, it allows them do exponentially more
ZLWK�WKDW�VDPH�VHW�RI�Ζ7�VHUYLFHV�UXQQLQJ�RQ�WRS�RI�WKH�VWRUDJH��6WHZDUW�DGGHG��
Ȋ$W�3XUH�6WRUDJH��ZH�FDQ�SURYLGH�D�VROXWLRQ�that can handle 10 times the capacity at
about the same cost of overhead as a spin-
QLQJ�GLVN�VROXWLRQ�ȋ�6WHZDUW�VDLG��Ȋ7KDW�DOORZV�agencies to keep up with the demand without
KDYLQJ�WR�FRPH�XS�ZLWK�DGGLWLRQDO�IXQGLQJ��:LWK�D�EXVLQHVV�PRGHO�FDOOHG�(YHUJUHHQ�6WRU-DJH��3XUH�RHUV�ȵDW�DQQXDO�PDLQWHQDQFH�DQG�allows agencies to stay modern by upgrading
software and controllers (free of charge)
HYHU\�WKUHH�\HDUV�ȋ��7KLV�PRGHO�JLYHV�DJHQFLHV�WUHPHQGRXV�ȵH[LELOLW\�DQG�SHDFH�RI�PLQG�
“Analytics brings to the table an improved
HQG�XVHU�H[SHULHQFH�EHFDXVH�LWȇV�D�SURDFWLYH�model for what is happening, trending or
changing, and how to immediately implement
D�FKDQJH�RU�DFWLRQ�ȋ�6WHZDUW�VDLG�
:LWK�3XUH�6WRUDJH��DJHQFLHV�QRW�RQO\�JHW�WKH�EHQHȴW�RI�ȵDVK�VWRUDJH�EXW�DOVR�VHDPOHVV�performance that is not impacted by mainte-
QDQFH�RU�HYHQ�XSJUDGHV���ΖWȇV�D�PXFK�VLPSOHU�(and less expensive) alternative to the indus-
try norm, where agencies have to rip and
replace storage regularly due to spectacular
LQFUHDVHV�LQ�PDLQWHQDQFH�FRVWV�
24 | A GOVLOOP GUIDE
Norcross Police DepartmentAn interview with Bill GroganSupport Services Captain at Norcross Police Department in Georgia
1R�SROLFH�GHSDUWPHQW�KDV�WKH�PDQSRZHU�WR�EH�HYHU\ZKHUH�DW�DOO�WLPHV��(YHQ�LQ�D�FLW\�WKH�VL]H�RI�1RUFURVV��*D���ZKLFK�VSDQV�RQO\�����VTXDUH�PLOHV��WKDWȇV�VWLOO�LPSRVVLEOH��
6RPH��������SHRSOH�FDOO�1RUFURVV�KRPH��EXW�QRERG\�NQRZV�KRZ�PDQ\�FRPH�LQWR�WKH�FLW\�HYHU\�GD\�IRU�ZRUN��Ȋ7KH�SUREOHP�WKDW�ZH�have is, for such a small area, we have quite a number of armed rob-
EHULHV�DQG�EXUJODULHV�ȋ�VDLG�%LOO�*URJDQ��&DSWDLQ�RI�6XSSRUW�6HUYLFHV�IRU�WKH�1RUFURVV�3ROLFH�'HSDUWPHQW��ȊΖ�GRQȇW�KDYH�WKH�UHVRXUFHV�WR�FRYHU�����VTXDUH�PLOHV�����KRXUV�D�GD\��VHYHQ�GD\V�D�ZHHN�ȋ
/LNH�PDQ\�RWKHU�ODZ�HQIRUFHPHQW�DJHQFLHV��WKH�GHSDUWPHQW�KDV�UHOLHG�RQ�KHDW�PDSV�WR�SLQSRLQW�KRW�VSRWV�IRU�FULPH��ȊΖWȇV�ROG�VFKRRO��LQ�P\�RSLQLRQ�ȋ�*URJDQ�VDLG��ȊΖWȇV�WRR�FRQYROXWHG�IRU�WKH�DYHUDJH�RɝFHU�WR�XQGHUVWDQG�ZKHUH�KH�RU�VKH�QHHGV�WR�EH�ȋ
3OXV��WKHVH�PDSV�FDQ�RQO\�JLYH�RɝFHUV�D�KLVWRULFDO�YLHZ�RI�ZKHUH�FULPH�KDV�RFFXUUHG��7KDW�NLQG�RI�GDWD�LV�ȴQH�IRU�YLHZLQJ�WUHQGV�RYHU�WLPH��
EXW�IRU�RɝFHUV�SDWUROOLQJ�WKH�VWUHHWV��WKH�VROXWLRQ�LV�UHDFWLYH�ȃ�QRW�SURDFWLYH��
Ȋ:H�ZDQWHG�WR�ȴQG�VRPHWKLQJ�WR�OHVVHQ�WKH�QXPEHU�RI�SURSHUW\�FULPHV�ȋ�*URJDQ�VDLG��%XW�WKH�GHSDUWPHQW�ZDVQȇW�VHDUFKLQJ�IRU�MXVW�DQ\�VROXWLRQ��ΖW�ZDQWHG�WR�LQYHVW�LQ�UHVRXUFHV�WKDW�FRXOG�SURDFWLYHO\�VKRZ�RɝFHUV�ZKHUH�FULPH�LV�PRVW�SUREDEOH�
In 2012, a new police chief took the reins and shared his vision of us-
LQJ�WHFKQRORJ\�WR�PDNH�HYHU\RQHȇV�MRE�HDVLHU��QRW�PRUH�FXPEHUVRPH��$ERXW�D�\HDU�ODWHU��ERWK�*URJDQ�DQG�3ROLFH�&KLHI�:DUUHQ�6XPPHUV�stumbled upon an article about the use of predictive analytics in law
HQIRUFHPHQW��
Ȋ:H�ZRQGHUHG�LI�LW�ZDV�SRVVLEOH�WR�SUHGLFW�FULPH�ȋ�*URJDQ�VDLG�
THE CHALLENGE: Predicting & Reducing Crime
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 25
1 2 3
7KDW�TXHVWLRQ�ZDV�SXW�WR�UHVW�RQ�WKH�ȴUVW�GD\�WKH�GHSDUWPHQW�UROOHG�RXW�SUHGLFWLYH�SROLFLQJ�VRIWZDUH�LQ�������2ɝFHUV�QDEEHG�VRPHRQH�wanted for burglary, and a few weeks later they caught burglars break-
LQJ�LQWR�D�KRPH��7KRVH�DUH�MXVW�D�IHZ�RI�WKH�VXFFHVV�VWRULHV�
7KLV�LV�KRZ�LW�ZRUNV��7KH�FRPSDQ\�WKDW�SURYLGHV�WKH�VRIWZDUH�JHWV�UHDO�WLPH�GDWD�IHHGV�DQG�KLVWRULFDO�GDWD�IURP�WKH�GHSDUWPHQWȇV�UHFRUGV�management system, including the type of crime, where it happened,
DQG�WKH�GDWH�DQG�WLPH�LW�RFFXUUHG��8VLQJ�DOJRULWKPV�DQG�YHU\�IDVW�GDWD�FUXQFKLQJ��WKH�VRIWZDUH�JHQHUDWHV�ȊSUHGLFWLRQ�ER[HVȋ�WKDW�VKRZ�RɝFHUV�WKH�PRVW�SUREDEOH�ORFDWLRQ�ZKHUH�D�FULPH�PLJKW�RFFXU��
Ȋ7KLV�LV�QRW�D�FU\VWDO�EDOO�RU�Ȇ0LQRULW\�5HSRUW�ȇȋ�VDLG�*URJDQ��UHIHUULQJ�WR�WKH�VFLHQFH�ȴFWLRQ�PRYLH�LQ�ZKLFK�SROLFH�XVHG�WHFKQRORJ\�WR�DUUHVW�PXUGHUHUV�EHIRUH�WKH\�FRPPLWWHG�FULPHV��
7KH�GDWD�WKDW�RɝFHUV�UHFHLYH�IURP�WKH�VRIWZDUH�WRRO�LV�LQ�UHDO�WLPH��:KHQ�WKH\�ȴUVW�FRPH�LQ�IRU�WKHLU����KRXU�VKLIW��RɝFHUV�ORJ�RQ�WR�WKHLU�computers to view the prediction boxes — eight red boxes, each
UHSUHVHQWLQJ�D�����E\�����IRRW�DUHD��VSUHDG�RYHU�D�PDS��2ɝFHUV�FDQ�zoom in on the boxes, which might encompass an apartment building,
D�JDV�VWDWLRQ�RU�D�URDGZD\��
$OO�RI�WKH�GHSDUWPHQWȇV�SROLFH�FDUV�DUH�HTXLSSHG�ZLWK�*36��DQG�VX-
SHUYLVRUV�FDQ�PRQLWRU�WKH�DPRXQW�RI�WLPH�WKDW�RɝFHUV�VSHQG�LQ�WKH�ER[HV��ΖI�RɝFHUV�VSHQG�WRR�PXFK�WLPH�LQ�WKH�ER[HV��WKH\ȇUH�QHJOHFW-LQJ�RWKHU�SDUWV�RI�WKH�FLW\��7KH�JRDO�LV�WR�VSHQG�D�OLWWOH�PRUH�WLPH�LQ�the prediction boxes than in other areas because that increases the
FKDQFH�RI�SUHYHQWLQJ�FULPH��*URJDQ�VDLG��7KRVH�ER[HV�FKDQJH�IRU�HY-HU\�RɝFHU��GHSHQGLQJ�RQ�WKH�WLPH�RI�GD\��WKH�PRVW�UHFHQW�FULPH�GDWD��DQG�WKH�W\SH�RI�FULPHV�WKH�GHSDUWPHQW�ZDQWV�WR�WDUJHW�
It took more than a year to get everyone on board with the software,
*URJDQ�VDLG��ΖW�ZDV�HDVLHU�WR�JHW�QHZ�RɝFHUV�WR�XVH�WKH�VRIWZDUH�EHFDXVH�WKH\�KDG�QR�RWKHU�URXWLQH�FRPLQJ�LQ�WR�WKH�GHSDUWPHQW��%XW�YHWHUDQ�RɝFHUV�ZHUH�DQRWKHU�VWRU\��Ȋ3ROLFH�RɝFHUV�GR�QRW�OLNH�FKDQJH��:H�SUHIHU�WKLQJV�JR�OLNH�WKH\ȇYH�EHHQ�JRLQJ�ȋ�KH�VDLG�
ΖQ�DGGLWLRQ�WR�WUDFNLQJ�WKH�H[WHQW�WR�ZKLFK�RɝFHUV�PRQLWRU�WKHLU�SUH-
GLFWLRQ�ER[HV��WKH�GHSDUWPHQW�DOVR�WUDFNV�ZKHWKHU�RU�QRW�RɝFHUV�XVH�WKH�VRIWZDUH��
*URJDQ�VDLG�RɝFHUV�KDYH�EHFRPH�LQFUHDVLQJO\�JRRG�DW�UHFRJQL]LQJ�DU-HDV�ZKHUH�WKHUHȇV�D�KLJKHU�SUREDELOLW\�RI�FULPH��7KH�WHFKQRORJ\�LV�HDV\�WR�XVH�DQG�UHTXLUHV�YHU\�OLWWOH�WUDLQLQJ��ZKLFK�KHOSHG�ERRVW�DGRSWLRQ�
SUCCESS FACTORS
THE SOLUTION
Properly train end users to use any new analytics software.
Seek their feedback on how to use the tool and improve processes.
Work with the data analytics provider to share valuable feedback on how to
improve the tool.
“This is not a crystal ball or ‘Minority Report.’”
26 | A GOVLOOP GUIDE
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 27
4 Keys to a Successful Data Analytics ProgramAn interview with Gus HornGlobal Consulting System Engineer, NetApp
INDUSTRY SPOTLIGHT
7KHUHȇV�QR�GHQ\LQJ�WKDW�ELJ�GDWD�DQDO\WLFV�has the power to transform the way govern-
PHQW�RSHUDWHV�DQG�VHUYHV�LWV�FLWL]HQV��7KHUH�are countless stories of agencies using ana-
lytics to speed emergency response times,
improve online services and better track
LQWHUQDO�PHWULFV��
But the path to a successful big data analytics
implementation is one that requires proper
long-term planning, said Gus Horn, Global
&RQVXOWLQJ�6\VWHP�(QJLQHHU�DW�1HW$SS�
Ȋ7KH�UHDVRQ�LV�WKHUH�DUH�PDQ\�YDULDEOHV�involved in acquiring and ingesting data into
ELJ�GDWD�DQDO\WLFV�SODWIRUPV��DQG�WKHQ�WKHUHȇV�also the logic around the analytic compo-
nent, how to write it and gain insight into
WKH�GDWD�ȋ�+RUQ�VDLG��Ȋ7KH�ELJJHVW�VWXPEOLQJ�block is typically around ingesting the data
LQWR�DQ�DQDO\WLF�SODWIRUP�ȋ
ΖWȇV�FRPPRQ�IRU�RUJDQL]DWLRQV�WR�VWDUW�RXW�with small development and testing environ-
ments and a handful of servers to demon-
VWUDWH�D�SURRI�RI�FRQFHSW��$OWKRXJK�WKHVH�small environments are a great starting point
for agencies to demonstrate the power of
big data analytics, the environment must
be adapted to properly plan for future data
JURZWK�DQG�ODUJHU�SURMHFWV���
Ȋ:KDW�ZH�ȴQG�LV�WKDW�WKHVH�YHU\�VPDOO�HQYLURQ-
PHQWV�DUH�H[WUHPHO\�SRZHUIXO�ȋ�+RUQ�VDLG��Ȋ6R�LW�FDQ�EH�YHU\�DOOXULQJ�IRU�RXU�FXVWRPHUV�worldwide, regardless of sector, to start down
WKDW�SDWK�DQG�VD\��Ȇ7KLV�ZRUNV�ZHOO��ΖȇP�MXVW�JR-
LQJ�WR�UHSHDW�DQG�GR�PRUH�RI�WKH�VDPH�ȇ�7KH\�become trapped in an architecture that is
extremely rigid in its design, and two or three
\HDUV�GRZQ�WKH�URDG�WKH\ȇUH�XQDEOH�WR�HYROYH��respond to newer technologies, or implement
QHZ�PHWKRGRORJLHV�LQ�WKHVH�SODWIRUPV�ȋ
7KH�SUREOHPV�W\SLFDOO\�GRQȇW�FRPH�XQWLO�ODWHU�on when the system has grown exponen-
tially, servers age and must be replaced,
DQG�GLVN�IDLOXUHV�DUH�D�FRQVWDQW�LVVXH��%\�WKDW�WLPH��LWȇV�YHU\�GLɝFXOW�WR�PRYH�WKH�GDWD�to a new hardware infrastructure, and the
employees who originally built the system
KDYH�ORQJ�OHIW�WKH�RUJDQL]DWLRQ��$W�WKDW�SRLQW��keeping the system up and running be-
FRPHV�YHU\�H[SHQVLYH��
7R�DYRLG�WKHVH�LVVXHV��+RUQ�UHFRPPHQGHG�WKDW�DJHQFLHV�SODQ�DFFRUGLQJO\�LQ�WKHVH�NH\�DUHDV��
Growth. Ȋ2UJDQL]DWLRQV�DOZD\V�KDYH�WR�SODQ�for growth when they start talking about a big
GDWD�SODWIRUP�ȋ�+RUQ�VDLG��Ȋ%LJ�GDWD�SODWIRUPV�by their nature are designed to accommo-
date the velocity, variety and variability of
FRQWHQW��DQG�EHFDXVH�RI�WKRVH�WKUHH�9ȇV�WKHVH�V\VWHPV�DUH�QHYHU�VWDWLF��6R�ZKLOH�DJHQFLHV�start out small, they should always be cogni-
]DQW�RI�WKH�IDFW�WKDW�WHFKQRORJ\�HYROYHV�ȋ�
7KH�FHQWUDO�SURFHVVLQJ�XQLWV�WKDW�DJHQFLHV�use are rapidly evolving and so are the
VWRUDJH�WHFKQRORJLHV��%XW�WKH�FKDOOHQJH�LV�these two technologies that play integral roles
in supporting big data systems are on two
GLHUHQW�JURZWK�SDWKV��+RUQ�VXJJHVWHG�WKDW�DJHQFLHV�ȴUVW�XQGHUVWDQG�WKH�SUREOHP�WKH\ȇUH�trying to address, as well as their growth
demands each month, each year and even
IXUWKHU�RXW�
Ȋ1HW$SSȇV�DUFKLWHFWXUH�LV�EDVHG�RQ�GLUHFWO\�attaching the storage to the compute, but
LW�JLYHV�DJHQFLHV�WKH�ȵH[LELOLW\�WR�GHFRXSOH�LW�and change out the computational element
without having to migrate or rebalance that
VWRUDJH�ȋ�+RUQ�VDLG��
)OH[LELOLW\��$JHQFLHV�WKDW�GRQȇW�FRQVLGHU�ȵH[L-ELOLW\�HDUO\�RQ�PD\�ȴQG�WKHPVHOYHV�ORFNHG�LQWR�
ZKDW�+RUQ�FDOOV�ȊULJLG�DUFKLWHFWXUDO�GHVLJQV�ȋ�ΖQ�WKLV�VFHQDULR��DJHQFLHV�ZLOO�ȴQG�LW�KDUGHU�WR�replace their servers when needed, without
KDYLQJ�WR�PLJUDWH�FRQWHQW��:LWK�1HW$SS��agencies can keep their data put when updat-
LQJ�WKHLU�WHFKQRORJ\�FOXVWHUV�
$JLOLW\��1RW�RQO\�DUH�ȵH[LEOH�FRPSXWH�DQG�storage important, but so is agility in provid-
LQJ�GLHUHQW�NLQGV�RI�VWRUDJH�VXEV\VWHPV��2QH�H[DPSOH�LV�WKH�DELOLW\�WR�PRYH�GLHUHQW�types of data to the most economical storage
options, depending on how often the data
LV�DFFHVVHG�DQG�LWV�LPSRUWDQFH��Ȋ7KHUHȇV�QR�SRLQW�LQ�NHHSLQJ�GDWD�WKDWȇV����\HDUV�ROG�DQG�not very relevant in your hot bucket, when
you can move that into a cold or glacier-like
storage device and still have access to it
VKRXOG�\RX�HYHU�QHHG�LW��<RX�FDQ�UHVHUYH�WKH�KLJK�SHUIRUPDQFH�ȵDVK�WHFKQRORJLHV�IRU�PRUH�PHDQLQJIXO�GDWD�ȋ�KH�VDLG�
2QH�RI�WKH�SHUNV�1HW$SS�SURYLGHV�LWV�customers is the ability to automate data
migration from higher-priced storage to
ORZHU�SULFHG�VWRUDJH��1HW$SS�ZRUNV�ZLWK�agencies to tag the data so they can track its
usage and determine which storage option
LV�PRVW�DSSURSULDWH��
(YROXWLRQ��Ȋ2IWHQWLPHV��DJHQFLHV�JHW�VWXFN�UH-
peating the same mistakes because the task
of moving to a newer technology is perceived
WR�EH�WRR�KHDY\�RI�D�OLIW�ȋ�+RUQ�VDLG��8VLQJ�1HW$SSȇV�GDWD�OLIHF\FOH�PDQDJHPHQW�WRRO��WKH�FRPSDQ\�FDQ�HDVLO\�LQWHJUDWH�LQWR�DQ�DJHQF\ȇV�existing technology cluster, migrate the data
to new hardware and decommission older
WHFKQRORJ\�JUDFHIXOO\��
7R�OHDUQ�PRUH��YLVLW��KWWS���ZZZ�QHWDSS�FRP�XV�VROXWLRQV�ELJ�GDWD�LQGH[�DVS[
28 | A GOVLOOP GUIDE
U.S. Postal ServiceAn interview with Dan Houston Jr.Manager of Data Science and Exploration, U.S. Postal Service
)HZ�RI�XV�VWRS�WR�FRQVLGHU�WKH�PDVVLYH�GDWD�DQDO\WLFV�RSHUDWLRQV�WKDW�KDSSHQ�EHKLQG�WKH�VFHQHV�HYHU\�WLPH�ZH�VHQG�DQG�UHFHLYH�PDLO��
ΖQ�D�VLQJOH�GD\��WKH�8�6��3RVWDO�6HUYLFH�SURFHVVHV�DQG�GHOLYHUV�PRUH�WKDQ�����PLOOLRQ�PDLOSLHFHV��7KDWȇV�PRUH�WKDQ���������PDLOSLHFHV�processed each minute, on average, or nearly 6,000 each second,
DQG�LWȇV�RQO\�D�VPDOO�VOLFH�RI�WKH�GDWD�8636�FROOHFWV�DQG�DQDO\]HV��7KH�agency relies on multiple data sets and tools to analyze operational
HɝFLHQFLHV�DQG�EHWWHU�SUHGLFW�ZKDW�UHVRXUFHV�DUH�QHHGHG�WR�PHHW�IXWXUH�GHPDQGV��
)RU�8636��WKH�SUREOHP�LVQȇW�D�ODFN�RI�GDWD��ΖQ�IDFW��8636�GDWD�LV�JURZLQJ�E\�DERXW����SHUFHQW�HDFK�\HDU��VDLG�'DQ�+RXVWRQ�-U���ZKR�VHUYHV�DV�WKH�DJHQF\ȇV�0DQDJHU�RI�'DWD�6FLHQFH�DQG�([SORUDWLRQ��ΖQ�WKDW�UROH��he oversees 35 petabytes of data — up from 22 petabytes in 2012 —
DORQJ�ZLWK�������GDWDEDVHV�DQG�����DSSOLFDWLRQV��
7KH�UHDO�LVVXH�LV�EHLQJ�DEOH�WR�TXLFNO\�SXOO�WRJHWKHU�ODUJH�GDWD�VHWV��DQDO\]H�WKHP�DQG�WKHQ�DQVZHU�EXVLQHVV�TXHVWLRQV��
“I started two years ago looking at gaps we had from a big data and an-
DO\WLFV�SHUVSHFWLYH�ȋ�+RXVWRQ�VDLG��Ȋ:H�KDYH�D�ORW�RI�WUDGLWLRQDO�WHFKQRO-ogies, a lot of traditional relational databases, and those work well to a
FHUWDLQ�SRLQW��%XW�ZKHQ�\RX�JHW�WR�WKH�ODUJHU�GDWD�VHWV�DQG�ZDQW�WR�EH�DEOH�WR�EXLOG�PRGHOV�DFURVV�\HDUV�RI�GDWD��WKH\�GRQȇW�ZRUN�YHU\�ZHOO�ȋ
'DWD�SURYLVLRQLQJ�ZDV�D�PDMRU�SDLQ�SRLQW��)RU�H[DPSOH��LI�VRPHRQH�ZDQWHG�WR�NQRZ�WKH�EHVW�PDUNHWLQJ�FDPSDLJQ�IRU�D�VSHFLȴF�WDUJHW�DXGLHQFH��RU�LQIRUPDWLRQ�DERXW�ZRUNORDGV�LQ�8636�PDLO�SURFHVVLQJ�plants, “it always turned into an eight- or nine-week data provisioning
H[HUFLVH�WR�EH�DEOH�WR�ZRUN�ZLWK�WKDW�GDWD�ȋ�+RXVWRQ�VDLG�
7KH�KDUG�SDUW�ZDV�JDWKHULQJ�WKH�GDWD�LQ�RQH�SODFH�VR�SHRSOH�FRXOG�work collectively to provide data-driven answers quickly, rather than
QLQH�ZHHNV�ODWHU��
%\�WKHQ��ȊZH�PLJKW�KDYH�PLVVHG�WKH�RSSRUWXQLW\�ȋ�+RXVWRQ�VDLG�
THE CHALLENGE: Improving Mail Delivery
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 29
1 2 3
7R�DGGUHVV�LWV�GDWD�SURYLVLRQLQJ�SUREOHPV��8636�WXUQHG�WR�WKH�RSHQ�VRXUFH�VRIWZDUH�+DGRRS��+RXVWRQ�VDLG�WKH�DJHQF\�ZDV�ORRNLQJ�IRU�D�platform that could store all types of data, and later format that data
ZKHQ�D�TXHU\�LV�PDGH�ȃ�D�SURFHVV�NQRZQ�DV�ODWH�ELQGLQJ��
But once the data is together in one place, how do you quickly visu-
DOL]H�WKH�GDWD�DQG�DFW�RQ�LW"�7KDWȇV�ZKHUH�WKH�DJHQF\ȇV�LQGH[LQJ�DQG�search tool, which integrates with the Hadoop environment, comes
LQWR�SOD\��
“By indexing and ingesting all of that data in one place, we can quickly
VHH�SDWWHUQV�WRR�ȋ�+RXVWRQ�VDLG��Ȋ6R�WKDWȇV�RQH�RI�WKH�JUHDW�SRZHUV�RI�DOO�RI�WKHVH�WRROV��1RW�RQO\�GR�\RX�KDYH�WKH�GDWD�LQ�RQH�SODFH�ȃ�DQG�you can certainly look at that raw data — but you can also quickly
JHW�YLVXDOL]DWLRQV��DQG�WKRVH�DUH�MXVW�RXW�RI�WKH�ER[��<RX�GRQȇW�QHHG�D�EXQFK�RI�GHYHORSHUV�WR�VWDUW�GHYHORSLQJ�GDVKERDUGV�ULJKW�DZD\��<RX�can get some quick visuals and explore the data quickly before you
PRYH�RQ�WR�GHYHORSLQJ�DQ�DSSOLFDWLRQ�ȋ�
8636�XVHV�WKHVH�YLVXDOL]DWLRQ�WRROV�LQ�YDULRXV�ZD\V��2Q�WKH�F\EHUVHFX-
ULW\�IURQW��WKH\�KHOS�WKH�DJHQF\�EHWWHU�XQGHUVWDQG�ZKHUH�LWȇV�SRWHQWLDOO\�H[SRVHG�WR�DWWDFNV��ΖWȇV�RQH�WKLQJ�WR�VD\�8636�WKZDUWHG��������DWWDFNV��EXW�LWȇV�IDU�PRUH�SRZHUIXO�ZKHQ�SHRSOH�FDQ�VHH�D�PDS�RI�ZKHUH�WKRVH�attacks originated and what parts of the environment attackers were
WU\LQJ�WR�HQWHU��
Ȋ7KH�RWKHU�ELJ�WKLQJ�ZHȇUH�ZRUNLQJ�RQ�ULJKW�QRZ�LV�SUHGLFWLYH�ZRUN-
loads, being able to better tell our plant managers and plant workers
QRW�MXVW�WKH�YROXPH�RI�SLHFHV�WKH\ȇUH�JRLQJ�WR�UHFHLYH��EXW�DOVR�ZKDW�equipment those pieces are going to have to be run through to get
through their facility — what pieces are going to go out to local post
RɝFHV��DQG�ZKDW�SLHFHV�DUH�JRLQJ�WR�JR�WR�DQRWKHU�SODQW�IRU�IXUWKHU�SURFHVVLQJ�ȋ�+RXVWRQ�VDLG�
Having this level of information will ideally help managers improve
HPSOR\HH�VFKHGXOLQJ��7KH�JRDO�LV�WR�SURYLGH�HPSOR\HHV�ZLWK�D����GD\�ZRUN�VFKHGXOH��UDWKHU�WKDQ�FKDQJLQJ�VFKHGXOHV�GDLO\��8VLQJ�WKH�GDWD��the agency can predict the path a mailpiece or package will take
through a plant and better address defects that may have thrown
WKDW�SDFNDJH�R�FRXUVH��VXFK�DV�PDLO�YROXPH��D�EURNHQ�PDFKLQH�RU�DQ�LVVXH�ZLWK�WKH�VRUWLQJ�HTXLSSLQJ��
'\QDPLF�URXWLQJ�LV�DQRWKHU�DUHD�RI�IRFXV�IRU�8636��7KH�JRDO�LV�WR�ȴQG�the optimal route for mail carriers, which may require adjusting routes
LQ�UHDO�WLPH�XVLQJ�*36�DQG�RWKHU�VHQVRU�GDWD��/HWȇV�VD\�RQH�FDUULHU�KDV�a big load and falls behind on some deliveries; dynamic routing would
allow the agency to shift some of that load to another carrier nearby
ZKR�KDV�D�OLJKWHU�ORDG�
SUCCESS FACTORS
THE SOLUTION
To reduce skepticism, find a pain point that’s relevant to people and show them how data can make it better.
Make data analytics tools and data accessible and usable to a wider au-
dience, not just people with advanced programming experience.
Focus on making data operational, so people can quickly visualize the infor-
mation and act on it.
“By indexing and ingesting all of that data in one place, we can quickly see patterns too. Not only do you have the data in one place but you can also quickly get visualizations, and those are just out of the box.”
USPS IN A DAY
512 $223.7320,132
money orders issued
113,531address changes processed
17,029passport applications accepted
millionmailpiecesprocessed & delivered
356,103a minute
5,935a second
million in revenue
30 | A GOVLOOP GUIDE
WebFOCUS iWay Software Omni
informationbuilders.com Get Social
Transforming Data Into Customer Value Customer-Facing Analytics
Differentiated User Experience
Information as an App
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 31
Data Analytics as an AssetAn interview with Dr. Rado KotorovChief Innovation Officer and Vice President, Information Builders
INDUSTRY SPOTLIGHT
'DWD�DQDO\WLFV�KDV�ORQJ�EHHQ�D�EX]]ZRUG�IRU�JRYHUQPHQW�DQG�WKH�SXEOLF�VHFWRU��ΖWȇV�XVHG�LQ�D�QXPEHU�RI�ZD\V�WR�EHWWHU�VHUYH�FLWL]HQV��But despite its growing popularity, govern-
ment continues to face many challenges with
GDWD�DQDO\WLFV��*LYHQ�WKH�YDVW�TXDQWLW\�RI�GDWD�out there and the cost of management and
analytical processes, many agencies have yet
WR�WUDQVIRUP�GDWD�DQDO\WLFV�LQWR�D�UHDO�DVVHW�
:K\�LV�WKLV"�)LUVW��JRYHUQPHQW�RUJDQL]DWLRQV�face a huge challenge when it comes to integra-
WLRQ�DQG�LQWHUFRQQHFWHGQHVV�RI�GDWD�VRXUFHV��'LHUHQW�GDWD�DUH�VWRUHG�LQ�GLHUHQW�GDWD�sources in a variety of formats, which makes
LW�HVSHFLDOO\�GLɝFXOW�DQG�H[SHQVLYH�WR�DFFHVV��WUDQVODWH�DQG�DQDO\]H��7KH�PRUH�YDULHW\�RI�GDWD�sources and formats, the more costly it is for
government to run analytics across various
RUJDQL]DWLRQV��
6HFRQG��WKHUHȇV�WKH�GLɝFXOW\�RI�SURSHUO\�SXE-
lishing and disseminating data to their con-
VWLWXHQWV�Ȃ�ZKHWKHU�LWȇV�WKH�SXEOLF��SDUWQHUV�RU�FXVWRPHUV��7UDQVIRUPLQJ�GDWD�WR�VHUYH�LQ�a public, self-service capacity requires new
WHFKQRORJLHV�DQG�QHZ��LQQRYDWLYH�IRUPDWV��When governments fail to provide accessible
data to their constituents, they lose account-
DELOLW\�DQG�WUDQVSDUHQF\�LQ�WKH�SXEOLF�H\H�
6R�KRZ�GRHV�JRYHUQPHQW�WDFNOH�WKHVH�challenges and use data analytics to turn
GDWD�LQWR�DQ�DVVHW"�'U��5DGR�.RWRURY��&KLHI�ΖQQRYDWLRQ�2ɝFHU�DQG�9LFH�3UHVLGHQW�RI�ΖQIRU-mation Builders, shared how government can
navigate these data analytics challenges and
KRZ�ΖQIRUPDWLRQ�%XLOGHUV�FDQ�KHOS��
Information Builders is in the business of
providing full information lifecycle management,
which includes the tools to access, manage,
DQDO\]H�DQG�SXEOLVK�GDWD��Ȋ:KHUH�ZH�KHOS�WKH�most is providing platforms to publish this infor-
PDWLRQ�WR�FLWL]HQV�DQG�HQG�XVHUV�ȋ�.RWRURY�VDLG��
7KHVH�SODWIRUPV�DOORZ�FRQVXPHUV�WR�LQWHUDFW�with the data, rearrange it, and customize it
without having to learn new tools or install
DQ\WKLQJ�QHZ�RQ�WKHLU�PDFKLQHV��
“We try to make it easier for organizations to
distribute analytical information to the end
XVHU�ZLWKRXW�UHTXLULQJ�DQ\�HRUW�RQ�WKH�XVHUȇV�SDUW�ȋ�.RWRURY�VDLG��
6RXQGV�OLNH�D�TXLFN�ȴ[�IRU�WKHVH�FKDOOHQJHV��but what makes data analytics more compli-
FDWHG�LV�WKDW�VROXWLRQV�RIWHQ�VLPSO\�FDQȇW�FRP-
PXQLFDWH�ZLWK�HDFK�RWKHU��.RWRURY�H[SODLQHG�WKH�GLɝFXOW\�RI�DFKLHYLQJ�LQWHUFRQQHFWHG�GDWD�through the concept of data warehouses
YHUVXV�GDWD�ODNHV��
Ȋ<RX�KDYH�LQIRUPDWLRQ�LQ�RQH�GDWD�ZDUHKRXVH�DQG�GLHUHQW�LQIRUPDWLRQ�LQ�DQRWKHU�GDWD�ZDUHKRXVH�WKDW�LV�SK\VLFDOO\�VHSDUDWHG�ȋ�KH�VDLG��Ȋ:LWK�D�SK\VLFDO�GDWD�ZDUHKRXVH��\RX�could have many individual data warehouses
and data marts and even many more Excel
VKHHWV�DQG�RWKHU�SHUVRQDO�GDWD�ȴOHV�ȋ�
7KHVH�SK\VLFDO�VLORHV�PDNH�LW�GLɝFXOW�WR�KDYH�seamless communication between an organi-
]DWLRQȇV�YDULRXV�OHYHOV�RI�GDWD��
7KDWȇV�ZK\�PDQ\�DUH�WXUQLQJ�WR�D�GDWD�ODNH��which is one big data storage pool that
KRXVHV�GLHUHQW�IRUPV�RI�GDWD��Ȋ7KH�GDWD�ODNH�is a logical concept, that is why some people
FDOO�LW�WKH�ORJLFDO�GDWD�ZDUHKRXVH�ȋ�.RWRURY�VDLG��Ȋ7KLV�DSSURDFK�KHOSV�\RX�HOLPLQDWH�WKH�VLORV��EXW�LW�UHTXLUHV�D�YHU\�ȵH[LEOH�LQWHJUDWLRQ�technology that allows you to quickly access
WKH�GLHUHQW�VHFWLRQV�RI�WKH�SRRO�RI�GDWD�ȋ
In this manner, the data lake becomes a logi-
cal structure that maps how you can connect
GLHUHQW�VRXUFHV�DQG�KRZ�\RX�FDQ�JHW�GDWD�
from each source and combine them togeth-
HU��7KDWȇV�ZK\�ΖQIRUPDWLRQ�%XLOGHUV�DOVR�KHOSV�organizations combine their data warehouses
LQWR�D�GDWD�ODNH��
&UHDWLQJ�GDWD�ODNHV��KRZHYHU��LV�RQO\�SDUW�RI�WKH�VROXWLRQ��.RWRURY�VD\V�JRYHUQPHQW�DOVR�needs to focus on taking the following steps
LQ�RUGHU�WR�IXOO\�XWLOL]H�GDWD�DV�DQ�DVVHW�
���ΖQWHJUDWH�GLHUHQW�GDWD�VRXUFHV�ZLWKLQ�GDWD�ODNHV��7KLV�HQDEOHV�\RXU�RUJDQL]DWLRQ�WR�get more in-depth and seamless information
IURP�\RXU�GDWD��
���0DLQWDLQ�WKH�GDWD��Maintaining your
data for data analytics is like maintaining the
TXDOLW\�RI�D�KRXVH��ΖW�QHHGV�WR�EH�URXWLQHO\�monitored, checked for problems, and re-
SDLUHG�DFFRUGLQJO\��7KLV�HQVXUHV�ORQJHU�GDWD�VKHOI�OLIH�
���$QDO\]H�DQG�SXEOLVK�\RXU�GDWD��2UJDQL]D-
tions need to ensure they have the analytical
FDSDELOLWLHV�WR�DQDO\]H�DQG�SXEOLVK�WKHLU�GDWD��:KHQ�VWULYLQJ�WRZDUG�WKH�ȴQDO�SURGXFW��DO-ZD\V�NHHS�WKH�HQG�XVHU�LQ�PLQG��0DQDJH�WKH�information to turn it into a product consum-
DEOH�E\�WKH�FLWL]HQ�RU�HQG�FXVWRPHU��
6WURQJHU�GDWD�DQDO\WLFV�LV�GRQH�EHVW�ZKHQ�FRQVLGHULQJ�WKH�XOWLPDWH�FRQVXPHU��7KDWȇV�why Information Builders works to provide
the tools and platforms necessary for govern-
ment to facilitate better analysis, integration,
SXEOLFDWLRQ�DQG�GLVVHPLQDWLRQ�RI�GDWD��7KLV�LV�critical to making data analytics more valuable
to government agencies, more accessible to
citizens, and ultimately, ensuring more trust
EHWZHHQ�JRYHUQPHQWV�DQG�WKHLU�FRQVWLWXHQWV�
:KDWȇV�PRVW�LPSRUWDQW�LV�WKDW�JRYHUQPHQW�continues to keep the consumer in mind
when transforming data analytics from a
IRUPLGDEOH�FKDOOHQJH�LQWR�D�YDOXDEOH�DVVHW�
32 | A GOVLOOP GUIDE
Pacific Northwest National LaboratoryAn interview with Justin BrownIT Engineer at Pacific Northwest National Laboratory
:KHQ�\RXȇUH�VXSSRUWLQJ�VFLHQWLȴF�WHDPV�IRFXVHG�RQ�SUHYHQWLQJ�DFWV�RI�terrorism and the non-proliferation of weapons of mass destruction,
\RX�KDYH�WR�JHW�LW�ULJKW��(QVXULQJ�WKRVH�UHVHDUFKHUV�KDYH�XQLQWHUUXSW-HG�DFFHVV�WR�IXQFWLRQDO�ZRUNVWDWLRQV�DQG�RQOLQH�UHVRXUFHV�LV�FULWLFDO��
Ȋ2XU�MRE�LV�MXVW�WR�GR�ZKDW�ZH�FDQ�WR�VXSSRUW�DQG�HQDEOH�WKDW�PLV-VLRQ�ȋ�VDLG�Ζ7�(QJLQHHU�-XVWLQ�%URZQ�RI�WKH�3DFLȴF�1RUWKZHVW�1DWLRQDO�/DERUDWRU\��311/���Ȋ:H�ZDQWHG�WR�JR�EH\RQG�MXVW�OLPLWLQJ�WKH�LPSDFW�RQ�FXVWRPHUV�ZKHQ�ZH�KDG�SUREOHPV��:H�ZDQWHG�WR�LPSURYH�WKH�customer experience because the more time that we save for our
customers — the researchers — the more it enables them to focus on
WKH�ODE�PLVVLRQ��Ȋ
7KH�FKDOOHQJH�IRU�%URZQ�DQG�KLV�WHDP�LV�WKDW�GLHUHQW�GHSDUWPHQWV�ZLWKLQ�WKH�ODUJHU�Ζ7�RUJDQL]DWLRQ�FROOHFW�GDWD�IRU�GLHUHQW�XVHUV��EXW�WKHUH�ZDVQȇW�DOZD\V�D�FRQVROLGDWHG�YLHZ�RI�WKDW�GDWD��
+DYLQJ�WKH�GDWD�WRJHWKHU�LQ�RQH�SODFH�KDV�LPSURYHG�RSHUDWLRQV��1RZ�HYHU\RQH�KDV�DFFHVV�WR�LW�DQG�FDQ�GR�FRUUHODWLRQV�DFURVV�GLHUHQW�GHSDUWPHQWV��%URZQ�VDLG��%XW�JHWWLQJ�WKHUH�ZDV�D�SURFHVV��
ΖW�LVQȇW�HDV\�WR�GR�WKDW�ZKHQ�HDFK�JURXS�ZLWKLQ�WKH�Ζ7�GHSDUWPHQW�KDV�LWV�RZQ�WRROV��)RU�H[DPSOH��WKH�QHWZRUN�WHDP�KDV�DQDO\WLFV�WRROV�IRU�monitoring its devices, the desktop team has tools for monitoring logs
and issues with the workstations, and the server administrators have
WRROV�DV�ZHOO��
ȊΖI�RQH�VHUYLFH�ZDV�DHFWLQJ�DQRWKHU��KDYLQJ�WKDW�YLVLELOLW\�RU�DFFHVV�WR�WKRVH�WRROV�ZDVQȇW�DOZD\V�WKHUH�ȋ�%URZQ�VDLG��
7KH�&Ζ2�SXW�LQ�SODFH�D�WHDP�WR�IRFXV�RQ�LPSURYLQJ�Ζ7�RSHUDWLRQV�WKURXJK�DXWRPDWLRQ�DQG�PRQLWRULQJ��3ULRU�WR�FUHDWLQJ�WKH�ȴYH�SHUVRQ�&KLHI�ΖQIUDVWUXFWXUH�$UFKLWHFWXUH�WHDP��WKHVH�HQJLQHHUV�ZHUH�ERJJHG�GRZQ�ZLWK�PRUH�UHSHWLWLYH�DQG�DGPLQLVWUDWLYH�WDVNV��
THE CHALLENGE: Enhancing IT Investment Tracking
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 33
1 2 3
311/�LV�QRZ�XVLQJ�GDWD�DQDO\WLFV�WR�JOHDQ�YDOXDEOH�IHHGEDFN�DERXW�WKH�RSHUDWLRQ�DQG�FRQGLWLRQ�RI�WKH�ODEȇV�Ζ7�VRXUFHV�DQG�KRZ�ZHOO�WKH\ȇUH�VHUYLQJ�HQG�XVHUV��%XW�WR�XQGHUVWDQG�KRZ�PXFK�GDWD�DQDO\WLFV�KDV�HQKDQFHG�Ζ7�RSHUDWLRQV�DW�311/��\RX�KDYH�WR�VWDUW�ZLWK�WKH�GDWD��
ΖQ�WKH�SDVW��WKH�Ζ7�SHUIRUPDQFH�GDWD�H[LVWHG�EXW�LW�ZDVQȇW�DFFHVVLEOH�GHSDUWPHQWZLGH��7KH�GLHUHQFH�WRGD\�LV�WKDW�ZKHQ�WKH�GHSDUWPHQW�monitors the performance and usability of servers, webpages and
RWKHU�Ζ7�DVVHWV��LW�FDQ�DOVR�VHDUFK�IRU�FRUUHODWLRQV�DQG�WUHQGV��
Having access to the data makes it easier to perform problem analy-
VHV��%URZQ�VDLG��1R�ORQJHU�GRHV�Ζ7�KDYH�WR�ORJ�LQ�WR�HDFK�VHUYHU�DQG�FRPE�WKURXJK�Ζ7�ORJV�WR�ȴJXUH�RXW�ZKDW�KDSSHQHG�DQG�ZKHUH�WKLQJV�ZHQW�ZURQJ��ΖI�WKHUHȇV�D�QHWZRUN�JOLWFK��WKH�GHSDUWPHQW�FDQ�HDVLO\�review its data for any anomalies or issues that occurred around the
VDPH�WLPH��
ΖWȇV�DOVR�HDVLHU�WR�VHH�LI�WKHUHȇV�D�VSLNH�LQ�DFFRXQW�ORFNRXWV��RU�VRIWZDUH�SDWFKHV�WKDW�DUHQȇW�LQVWDOOLQJ�SURSHUO\��Ȋ:H�FDQ�VWRS�WKRVH�SUREOHPV�EHIRUH�WKH\�JHW�WRR�ELJ�ȋ�%URZQ�VDLG��
$V�WKH�GHSDUWPHQWȇV�PRQLWRULQJ�DQG�DXWRPDWLRQ�FDSDELOLWLHV�LPSURYH��FXVWRPHU�VHUYLFH�UHPDLQV�DW�WKH�IRUHIURQW��)RU�LQVWDQFH��UDWKHU�WKDQ�HPDLOLQJ�WKH�ZKROH�ODE�DERXW�D�ZHEVLWH�WKDWȇV�JRLQJ�WR�EH�RɞLQH��Brown and his team can use more targeted communications to inform
WKH�VSHFLȴF�XVHUV�RI�WKH�VLWH��
7KH�WHDP�FDQ�XVH�GDWD�WR�GHWHUPLQH�SHDN�XVH�WLPHV�DQG�VFKHGXOH�PDLQWHQDQFH�DURXQG�WKDW��%XW�WKH�XVHU�H[SHULHQFH�GRHVQȇW�VWRS�WKHUH��
7KH�GDWD�DOVR�VKRZV�KRZ�PDQ\�FOLFNV�LW�WDNHV�XVHUV�WR�ȴQG�ZKDW�WKH\ȇUH�ORRNLQJ�IRU�RQOLQH��%URZQ�VHHV�WKLV�LQVLJKW�VKDSLQJ�XVHU�LQWHUIDFH�LPSURYHPHQWV�LQ�WKH�IXWXUH�
Ζ7�SXUFKDVLQJ�LV�DQRWKHU�DUHD�ULSH�IRU�GDWD�DQDO\WLFV�DW�311/��+HUHȇV�ZK\��8VLQJ�GDWD�DERXW�WKH�DJH�RI�D�V\VWHP��KRZ�LWȇV�SHUIRUPLQJ��KRZ�LWȇV�ORDGLQJ��DQG�KRZ�RIWHQ�LWȇV�XVHG��WKH�GHSDUWPHQW�FDQ�DVVLJQ�HDFK�RI�LWV�Ζ7�UHVRXUFHV�D�ULVN�VFRUH��7KLV�DOORZV�WKH�WHDP�WR�SURDFWLYHO\�GHWHUPLQH�ZKDW�VKRXOG�EH�UHIUHVKHG�RU�ZKHUH�DGGLWLRQDO�UHVRXUFHV�DUH�QHHGHG��
ȊΖQ�WKH�SDVW��SUREDEO\�RXU�ELJJHVW�QRWLȴFDWLRQ�RI�D�SUREOHP�ZDV�ZKHQ�FXVWRPHUV�FDOOHG�WKH�KHOS�GHVN�ȋ�%URZQ�VDLG��ȊΖI�ZH�KDYH�WR�ZDLW�IRU�WKDW�QRWLȴFDWLRQ��WKHQ�ZHȇYH�IDLOHG�ȋ
&XVWRPHUV�DUHQȇW�WKH�RQO\�RQHV�ZKR�EHQHȴW�IURP�WKH�GHSDUWPHQWȇV�LQYHVWPHQWV��
ΖQ�ȴVFDO�������%URZQ�HVWLPDWHV�WKDW�WKLV�GHSDUWPHQW�VDYHG�DERXW�������KRXUV�LQ�HɝFLHQFLHV�DQG����������WKURXJK�WKH�XVH�RI�GDWD�DQD-
O\WLFV�DQG�SURFHVV�LPSURYHPHQWV��
When Brown and his team began rolling out data analytics capabilities
about two years ago, the focus was on empowering cyberteams to
VHDUFK�WKURXJK�Ζ7�ORJV�IDVWHU��
Ȋ%XW�WKHQ��RQFH�ZH�VWDUWHG�VHHLQJ�WKH�SRWHQWLDO�RI�WKDW�GDWD�IRU�WKH�Ζ7�support side of things and the operational side of things, we just start-
HG�KHDYLO\�GLJJLQJ�LQWR�LW�DQG�VHHLQJ�ZKDW�ZH�FRXOG�GR�ȋ�%URZQ�VDLG��
SUCCESS FACTORS
THE SOLUTION
Determine key performance indicators, or what metrics will be measured. For each service, determine what matters
most to the organization and begin monitoring those issues.
Develop a methodology or a systematic approach to using the data. Without
that, having all the data won’t do a lot of good.
Show the return on investment. Use the data to clearly show time savings, cost
savings and any other benefits.
“Once we started seeing the potential of that data for the IT support side of things and the operational side of things, we just started heavily digging into it and seeing what we could do.”
accessible IT performance
data
improved monitor-ing & automation
capabilities
assigned risk score to IT resources
1,800 hours + $453,000
SAVED+ =+
34 | A GOVLOOP GUIDE
1 2 3
3 REASONS PRIVACY SHOULD BE TOP OF MIND ΖQ�0D\�������3UHVLGHQW�2EDPD�FKDUJHG�WKH�&RXQVHORU�WR�WKH�3UHVLGHQW�WR�OHDG�D����GD\�review�RI�GDWD�DQG�SULYDF\��7KH�SUHVLGHQW�ZDQWHG�WR�NQRZ�ZKDW�QHZ�WHFKQRORJLHV�ZHUH�GHȴQLQJ�WKH�GDWD�ODQGVFDSH�DQG�KRZ�ELJ�GDWD�DQG�GDWD�DQDO\WLFV�LPSDFW�SXEOLF�SROLF\�DQG�ODZV�WKDW�JRYHUQ�SULYDF\��SDUWLFXODUO\�IRU�FLWL]HQV��
7KH�ȴQGLQJV�IURP�WKDW�UHYLHZ�XVH�WKH�SKUDVHV�ȊELJ�GDWD�WRROVȋ�DQG�ȊGDWD�DQDO\WLFV�ȋ�EXW�GRQȇW�JHW�WRR�FDXJKW�XS�LQ�WKH�WHUPLQRORJ\��<HV��LWȇV�LPSRUWDQW��%XW�DV�ZH�QRWHG�HDUOLHU��WKH�WHUPLQRORJ\�FKDQJHV�IURP�SHUVRQ�WR�SHUVRQ��:KDWȇV�PRVW�LPSRUWDQW�WR�XQGHUVWDQG�LV�WKDW�GDWD�SULYDF\�GRHVQȇW�VWDUW�RU�VWRS�DW�D�VLQJOH�SRLQW�LQ�WLPH��3ULYDF\�KDV�WR�EH�WRS�RI�PLQG�EH-IRUH�WKH�GDWD�LV�HYHQ�FUHDWHG��DQG�LW�PXVW�EH�PDLQWDLQHG�ZKHQ�WKH�GDWD�LV�FROOHFWHG��VWRUHG��VKDUHG�DQG�DQDO\]HG��
$V�WKH�UHSRUW�QRWHV��Ȋ7KH�RSSRUWXQLWLHV�SUHVHQWHG�E\�ELJ�GDWD�DUH�FRQVLGHUDEOH��EXW�ELJ�GDWD�UDLVHV�VHULRXV�FRQFHUQV�DERXW�KRZ�ZH�SURWHFW�RXU�SULYDF\�DQG�RWKHU�YDOXHV�ȋ
ΖWȇV�WUXH�WKDW�ȊELJ�GDWD�GULYHV�ELJ�EHQHȴWV��IURP�LQQRYDWLYH�EXVLQHVVHV�WR�QHZ�ZD\V�WR�WUHDW�GLVHDVHV�ȋ�DFFRUGLQJ�WR�WKH�UHSRUW��%XW�ȊWKH�FKDOOHQJHV�WR�SULYDF\�DULVH�EHFDXVH�WHFKQRORJLHV�FROOHFW�VR�PXFK�GDWD��H�J���IURP�VHQVRUV�LQ�HYHU\WKLQJ�IURP�SKRQHV�WR�SDUNLQJ�ORWV��DQG�DQDO\]H�WKHP�VR�HɝFLHQWO\��H�J���WKURXJK�GDWD�PLQLQJ�DQG�RWKHU�NLQGV�RI�DQDO\WLFV��WKDW�LW�LV�SRVVLEOH�WR�OHDUQ�IDU�PRUH�WKDQ�PRVW�SHRSOH�KDG�DQWLFLSDWHG�RU�FDQ�DQWLFLSDWH�JLYHQ�FRQWLQXLQJ�SURJUHVV�ȋ
%HORZ�DUH�VRPH�FULWLFDO�SULYDF\�FRQVLGHUDWLRQV�IURP�WKH�ELJ�GDWD�UHSRUW�WKDW�\RX�VKRXOG�KHHG�RQ�\RXU�ELJ�GDWD�DQG�DQDO\WLFV�MRXUQH\�
Big data and analytics tools can alter the balance of power between government and citizen.
Government agencies can reap enormous
EHQHȴWV�IURP�XVLQJ�ELJ�GDWD�WR�LPSURYH�VHUYLFH�GHOLYHU\�RU�GHWHFW�SD\PHQW�IUDXG��%XW�government uses of big data also have the
potential to chill the exercise of free speech or
IUHH�DVVRFLDWLRQ��$V�PRUH�GDWD�LV�FROOHFWHG��DQ-
alyzed, and stored on both public and private
systems, we must be vigilant in ensuring that
balance is maintained between government
DQG�FLWL]HQV��DQG�UHYLVH�RXU�ODZV�DFFRUGLQJO\�
Big data and analytics tools can reveal intimate personal details.
2QH�SRZHUIXO�ELJ�GDWD�WHFKQLTXH�LQYROYHV�merging multiple data sets, drawn from dis-
SDUDWH�VRXUFHV��WR�UHYHDO�FRPSOH[�SDWWHUQV��But this practice, sometimes known as “data
IXVLRQ�ȋ�FDQ�DOVR�OHDG�WR�WKH�VR�FDOOHG�PRVDLF�HHFW��ZKHUHE\�SHUVRQDOO\�LGHQWLȴDEOH�LQIRU-mation can be discerned even from ostensi-
EO\�DQRQ\PL]HG�GDWD��$V�ELJ�GDWD�EHFRPHV�more widely used in the private sector
to bring a wellspring of innovations and
SURGXFWLYLW\��ZH�PXVW�HQVXUH�WKDW�HHFWLYH�consumer privacy protections are in place to
SURWHFW�LQGLYLGXDOV�
Big data and analytics tools could lead to discriminatory outcomes.
As more decisions about our commercial and
personal lives are determined by algorithms
and automated processes, we must pay care-
ful attention that big data does not system-
atically disadvantage certain groups, whether
LQDGYHUWHQWO\�RU�LQWHQWLRQDOO\��:H�PXVW�SUH-
vent new modes of discrimination that some
uses of big data may enable, particularly with
regard to longstanding civil rights protections
LQ�KRXVLQJ��HPSOR\PHQW��DQG�FUHGLW�
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 35
WHAT YOU CAN DO NOW: HOW TO CREATE A CULTURE FOR DATA ANALYTICS$VLGH�IURP�EXGJHWV��FXOWXUDO�EDUULHUV�SRVH�RQH�RI�WKH�JUHDWHVW�FKDOOHQJHV�WR�LQWURGXF-LQJ�QHZ�WHFKQRORJLHV�DQG�QHZ�ZD\V�RI�WKLQNLQJ�LQWR�DQ\�DJHQF\��7KHUHȇV�D�FHUWDLQ�ZD\�WKDW�SHRSOH�KDYH�DOZD\V�GRQH�WKLQJV��DQG�VRPHWLPHV�OHDUQLQJ�D�QHZ�ZD\�FDQ�FDXVH�DQJVW�RU�SXVKEDFN��
6R�KRZ�GR�\RX�IRVWHU�DQ�HQYLURQPHQW�ZKHUH�SHRSOH�DUH�RSHQ�DQG�H[FLWHG�DERXW�WHVWLQJ�WKH�ZDWHUV�DQG�SXWWLQJ�GDWD�DQDO\WLFV�WR�XVH"�0D\EH�\RXȇUH�ZHOO�LQWR�\RXU�DQDO\WLFV�MRXUQH\��EXW�\RXȇYH�KLW�D�IHZ�VQDJV�DQG�QHHG�VRPH�DGYLFH�IURP�\RXU�SHHUV�ȃ�RU�PD\EH�\RXȇUH�MXVW�EHJLQQLQJ�WKDW�MRXUQH\��
:KHUHYHU�\RX�DUH�RQ�WKDW�SDWK��\RX�FDQ�EHQHȴW�IURP�WKH�OHVVRQV�OHDUQHG�E\�RWKHUV��:H�VSRNH�ZLWK�GDWD�H[SHUWV�DW�YDULRXV�JRYHUQPHQW�DJHQFLHV�WR�JOHDQ�SUDFWLFDO�WLSV�IRU�FUHDW-LQJ�D�FXOWXUH�IRU�GDWD�DQDO\WLFV�
Jeff Chen Chief Data Scientist,
Department of Commerce
Learn the lingo. When Chen was Director of Analytics for the New York City Fire Department, he rode DORQJ�ZLWK�ȴUHȴJKWHUV�WR�XQGHUVWDQG�WKHLU�GDLO\�H[SHULHQFHV�DQG�VWUXJJOHV��
Tell the story about your data. There are a lot of scientists at Commerce, DQG�WKH\�KDYH�D�ORW�RI�GDWD��7R�KHOS�FUHDWH�LQWHUHVW�LQ�WKLV�GDWD��WKH�DJHQF\�EXLOW�WXWRULDOV�WKDW�IHDWXUH�WKH�VFLHQWLVWV�WHOOLQJ�WKHLU�GDWD�VWRULHV��ȊΖWȇV�DOZD\V�D�GLDORJXH�ȋ�&KHQ�VDLG��
Jon Minkoff Chief Data Officer,
Enforcement Bureau, Federal Communications Commission
Brand your systems. $W�WKH�(QIRUFHPHQW�%XUHDX��HPSOR\HHV�know their reports will be accepted only if WKH\�FRPH�IURP�WKH�EXUHDXȇV�DXGLW�WUDFNLQJ�V\VWHP��7KDW�PHDQV�UHSRUWV�EDVHG�RQ�GDWD�IURP�VRPHRQHȇV�([FHO�VSUHDGVKHHW�ZRQȇW�FXW�LW��ȊΖW�PD\�EH�DFFXUDWH��EXW�LWȇV�QRW�WKH�URRW�ȋ�0LQNR�VDLG��%\�GHPDQGLQJ�WKDW�DOO�UHSRUWV�FRPH�IURP�WKH�EXUHDXȇV�V\VWHP��LW�IRUFHV�SHRSOH�WR�XQGHUVWDQG�WKH�V\VWHP�EHWWHU��DQG�LW�KHOSV�WKH�EXUHDX�LGHQWLI\�DQG�UHSDLU�JDSV��
Kirk Borne Data Scientist and former Professor of Astrophysics and Computational Science at George Mason University
Understand data as a literacy. While technical skills are important, they are RQO\�WUXO\�YDOXDEOH�ZKHQ�FRPELQHG�ZLWK�DQ�XQGHUVWDQGLQJ�RI�GDWD�DQG�SUREOHP�VROYLQJ��ȊΖ�DOZD\V�WHOO�P\�VWXGHQWV�WR�XQGHUVWDQG�GDWD�DV�D�OLWHUDF\�ȋ�%RUQH�VDLG�DW�D�0DUFK������*RY/RRS�HYHQW��Ȋ/LWHUDF\�LV�NQRZLQJ�KRZ�WR�XVH�GDWD��ZKHQ�WR�XVH�LW��ZKDW�W\SHV�WKHUH�DUH��DQG�ZKDW�W\SHV�RI�TXHVWLRQV�LW�FDQ�DQVZHU�ȋ
Recognize that data is an asset. 'RLQJ�VR�JLYHV�GDWD�D�SURPLQHQW�UROH��Ȋ'DWD�is a binary representation of real information, and real information is what we are trying WR�H[WUDFW�ȋ�%RUQH�VDLG��<RX�FDQ�XVH�GDWD�WR�create better insights, decisions, services to FLWL]HQV�RU�RWKHU�DJHQFLHV��
36 | A GOVLOOP GUIDE
Caryl Brzymialkiewicz Assistant Inspector General and Chief
Data Officer at the Department of Health and Human Services Office of
the Inspector General
Go on a listening tour. :KDW�EHWWHU�ZD\�WR�ȴQG�RXW�ZKDW�SHRSOH�DUH�WKLQNLQJ�WKDQ�DFWXDOO\�WDONLQJ�WR�WKHP"�7KH\�PD\�QRW�EH�GRZQ�WKH�KDOO��EXW�IROORZ�%U]\PLDONLHZLF]ȇV�OHDG��6KH�VFKHGXOHG�FRHH�PHHWLQJV�DQG�OXQFKHV�ZLWK�SHRSOH�RQVLWH�DQG�KHOG�6N\SH�FDOOV�ZLWK�SHRSOH�LQ�WKH�ȴHOG��$VN�WKHP�DERXW�WKHLU�WHQVLRQ�SRLQWV��ZKDW�WKH\�FDQȇW�DFFHVV�EXW�ZDQW�WR�DFFHVV��DQG�ZKDW�ZRXOG�KHOS�WKHP�GR�WKHLU�MREV�EHWWHU��
Consider the talent you have on board. 7KLQN�DERXW�KRZ�WR�EOHQG�IHGHUDO�KLULQJ�ZLWK�TXLFN��FRPPHUFLDO�R�WKH�VKHOI�VRIWZDUH��7KLQN�DERXW�OHYHUDJLQJ�3UHVLGHQWLDO�0DQDJH�PHQW�)HOORZV�RU�D�FROOHJH�LQWHUQ�WR�KHOS�\RX�EXLOG�\RXU�GDWD�DQDO\WLFV�WHDP��Ȋ+DYLQJ�RQH�VWXGHQW�RQ�ERDUG�FDQ�EULQJ�LQ�QHZ�WDOHQW�DQG�WKRXJKWV�ȋ�%U]\PLDONLHZLF]�VDLG�
Consider rotation programs. ΖV�WKHUH�DQ�HQWUHSUHQHXU�LQ�UHVLGHQFH�RU�DQ�H[HFXWLYH�WKDW�\RX�FDQ�EULQJ�RQ�ERDUG�WR�VXSSRUW�\RXU�HRUWV"�
Joshua Martin Chief of Staff,
the Indiana Management and Performance Hub
Make room at the table for a variety of participants. 7KH�WHDP�DW�03+�LV�ZRUNLQJ�ZLWK�VWDWH�DJHQFLHV�WR�H[SORUH�ZKDW�TXHVWLRQV�FDQ�EH�DQVZHUHG�XVLQJ�WKHLU�GDWD�DQG�WKH�VWDWHȇV�EXVLQHVV�LQWHOOLJHQFH�WRRO��6WDWH�agencies are a part of the planning process and will help shape how the tool LV�XVHG��Ȋ:H�UHDOO\�RSHQHG�WKLV�XS�WR�WKH�RSSRUWXQLW\�WR�FRQWLQXH�WR�EXLOG�RQ�LW�ZLWK�D�QXPEHU�RI�GLHUHQW�VWDNHKROGHUV�DQG�WR�EULQJ�HYHU\RQH�WR�WKH�WDEOH�ȋ�0DUWLQ�VDLG��
Pull from your past experiences. Before starting his career as a state govern�PHQW�HPSOR\HH��0DUWLQ�VSHQW�VL[�\HDUV�LQ�ȴUH�DQG�(06��+LV�SXEOLF�VDIHW\�EDFNJURXQG�DOORZV�KLP�WR�EHWWHU�FRPPXQLFDWH�DQG�UH�late to agencies that have a role in combat�LQJ�ΖQGLDQDȇV�RSLRLG�HSLGHPLF��
Keep everyone informed. 0DUWLQ�DQG�KLV�FROOHDJXHV�KDYH�HLWKHU�PHW�ZLWK�RU�DUH�VFKHGXOHG�WR�PHHW�ZLWK�HYHU\�DJHQF\�WKDW�LV�SURYLGLQJ�GDWD�WR�03+��7KH�JRDO�LV�WR�VKRZ�WKHP�ZKDW�03+�LV�GRLQJ�ZLWK�WKHLU�GDWD�DQG�KHOS�WKHP�WR�EHWWHU�XWL�OL]H�WKHLU�GDWD�WR�PDNH�GHFLVLRQV��Ȋ2QFH�\RX�FDQ�VKRZ�\RXU�SDUWQHUV�ZKDW�\RXȇUH�GRLQJ�DQG�ZKDW�YDOXH�\RX�FDQ�H[WUDFW�IURP�WKHLU�GDWD��SDUWLFLSDWLRQ�XVXDOO\�SLFNV�XS�IDLUO\�UDSLGO\�ȋ�0DUWLQ�VDLG��
Linda Powell Chief Data Officer,
Consumer Financial Protection Bureau
Your data scientists and data manage-ment professionals have to understand the jargon of the business. ȊΖI�Ζ�ZDON�LQWR�DQ�HFRQRPLVWȇV�RɝFH�DQG�Ζ�GRQȇW�XQGHUVWDQG�KLP�ZKHQ�KH�LV�WDONLQJ�DERXW�GR�ing a linear regression, he is going to kick me RXW�ȋ�3RZHOO�VDLG��<RX�KDYH�WR�XQGHUVWDQG�WKH�EXVLQHVV�DQG�WDON�WR�WKHP�LQ�WKHLU�ODQJXDJH��QRW�LQ�JLJDE\WHV�DQG�WHUDE\WHV�RI�GDWD��
Understand what your agency customers or partners want to accomplish. ΖWȇV�QRW�DV�VLPSOH�DV�ZDONLQJ�LQWR�VRPHRQHȇV�RɝFH�DQG�DVNLQJ�WKHP�ZKDW�WKH\�ZDQW��6RPHWLPHV�WKH\�GRQȇW�NQRZ��DQG�LW�PD\�EH�KDUG�IRU�WKHP�WR�DUWLFXODWH�LW��+HUHȇV�ZKHUH�\RX�QHHG�WR�VLW�GRZQ�DQG�KDYH�D�WZR�ZD\�FRQYHUVDWLRQ��)URP�WKDW�GLDORJXH��\RXȇOO�EH�VXUSULVHG�DW�KRZ�PXFK�\RX�OHDUQ�DERXW�WKHLU�GDWD�DQDO\WLFV�QHHGV���ΖI�\RX�GRQȇW�NQRZ�ZKDW�WR�DVN��FKHFN�RXW�WLSV�IURP�&DU\O�%U]\PLDON�LHZLF]��6KH�SURYLGHV�VRPH�VHHG�TXHVWLRQV��
HOW YOU CAN USE DATA ANALYTICS TO CHANGE GOVERNMENT | 37
ABOUT & ACKNOWLEDGMENTS
CONCLUSION'DWD�LV�SRZHUIXO��:H�FDQ�XVH�LW�WR�VKDSH�SROLFLHV��GHWHFW�IUDXG��ZDVWH�DQG�DEXVH��DQG�HYHQ�VDYH�OLYHV��%XW�LW�WDNHV�PRUH�WKDQ�GDWD�DORQH�WR�GULYH�EHWWHU�GHFLVLRQ�PDNLQJ�DQG�XOWLPDWHO\�EHW-ter outcomes.
:H�DOVR�QHHG�WKH�ULJKW�WRROV�WR�FRPELQH�WKDW�GDWD�DQG�VHDUFK�IRU�SDWWHUQV��DQRPDOLHV�DQG�WUHQGV�WKDW�RWKHUZLVH�ZRXOG�JR�XQGH-WHFWHG��7KLQN�EDFN�WR�WKH�0LFKLJDQ�RQFRORJLVW�ZKR�LQWHQWLRQDOO\�PLVGLDJQRVHG�KLV�SDWLHQWV��ΖPDJLQH�KRZ�PDQ\�PRUH�OLYHV�FRXOG�KDYH�EHHQ�GHVWUR\HG�ZLWKRXW�WKH�LQFULPLQDWLQJ�GDWD�WR�EDFN�XS�WKH�ZKLVWOHEORZHUȇV�FODLPV��
:H�KDYH�WR�VWRS�YLHZLQJ�GDWD�DQDO\WLFV�DV�VROHO\�DQ�Ζ7�SURMHFW�RU�VRPHWKLQJ�IRU�SHRSOH�ZLWK�FRPSXWHU�VFLHQFH�GHJUHHV��6RPH�RI�WKH�EHVW�GDWD�DQDO\WLFV�WHDPV�DUH�PXOWLGLVFLSOLQDU\��)RU�H[DPSOH��WKH�SHRSOH�RQ�*6$ȇV�+5�GDWD�DQDO\WLFV�WHDP�FRPH�IURP�GLYHUVH�ȴHOGV��ZLWK�EDFNJURXQGV�LQ�MRXUQDOLVP��ȴQDQFH��DQG�LQGXVWULDO�DQG�RUJDQL]DWLRQDO�SV\FKRORJ\��WR�QDPH�D�IHZ��
$V�GDWD�DQDO\WLFV�FRQWLQXHV�WR�SHUPHDWH�JRYHUQPHQW��WDNH�WKH�WLPH�WR�LGHQWLI\�SHRSOH�ZKR�FDQ�DGG�YDOXH�WR�\RXU�DQDO\WLFV�SURM-HFWV��)LQG�WKH�ULJKW�SHRSOH�WR�WUDLQ��DQG�LQYHVW�LQ�WKHP��<RX�QHYHU�NQRZ�ȃ�\RXU�EHVW�GDWD�VFLHQWLVW�PD\�EH�D�SKLORVRSK\�PDMRU��$IWHU�DOO��WKH\�DUH�JRRG�WKLQNHUV��
So what’s next for data analytics in government?
([SHUWV�LQ�WKH�ȴHOG�H[SHFW�WR�VHH�JUHDWHU�XVH�RI�GDWD�DQDO\WLFV�to drive cybersecurity operations and predictions about future WKUHDWV��:KHQ�DJHQFLHV�DUH�UHDFWLYH�ȃ�UDWKHU�WKDQ�SURDFWLYH�ȃ�WKH�RXWFRPHV�FDQ�EH�GHYDVWDWLQJ��2QH�UHFHQW�H[DPSOH�LV�WKH�PDVVLYH�GDWD�EUHDFK�DJDLQVW�2ɝFH�RI�3HUVRQQHO�0DQDJHPHQW�systems�WKDW�OHIW�PLOOLRQV�RI�VHQVLWLYH�HPSOR\HH�UHFRUGV�H[SRVHG�WR�KDFNHUV��
Ȋ&\EHU�LV�WKH�1R����LVVXH�RI�RXU�WLPH�ȋ�VDLG�%RUQH��ȊΖWȇV�QRW�MXVW�230��EXW�VR�PDQ\�LQVWDQFHV�ZKHUH�LW�ZRXOG�EH�EHWWHU�WR�SUHGLFW�something that doesn’t happen than not to predict anything and GHDO�ZLWK�WKH�IDOORXW�ȋ
6RPH�DJHQFLHV�DUH�H[SORULQJ�QH[W�VWHSV�IRU�XVLQJ�DQDO\WLFV�WR�EH�SURDFWLYH��UDWKHU�WKDQ�UHDFWLYH��ZKHQ�LW�FRPHV�WR�VHFXULW\�DQG�FDUU\LQJ�RXW�WKHLU�UHVSHFWLYH�PLVVLRQV��0DQ\�RI�WKH�DJHQFLHV�KLJK-OLJKWHG�LQ�RXU�FDVH�VWXGLHV�DUH�H[SORULQJ�WKRVH�RSWLRQV��LQFOXGLQJ�WKH�VWDWH�RI�ΖQGLDQD��*6$��WKH�++6�ΖQVSHFWRU�*HQHUDOȇV�2ɝFH�DQG�WKH�1RUFURVV�3ROLFH�'HSDUWPHQW��
.HHS�LQ�PLQG�WKDW�WKLV�UHVRXUFH�JXLGH�LV�PRUH�WKDQ�D�FRPSLODWLRQ�of government success stories. It’s an invitation to reach out to WKH�DJHQFLHV�\RXȇYH�UHDG�DERXW��DQG�D�OHDUQLQJ�WRRO�WR�HQKDQFH�\RXU�GDWD�DQDO\WLFV�MRXUQH\��
GRY/RRSȇV�PLVVLRQ�LV�WR�ȊFRQQHFW�JRYHUQPHQW�WR�LPSURYH�JRYHUQ-PHQW�ȋ�:H�DLP�WR�LQVSLUH�SXEOLF�VHFWRU�SURIHVVLRQDOV�E\�VHUYLQJ�DV�WKH�NQRZOHGJH�QHWZRUN�IRU�JRYHUQPHQW��*RY/RRS�FRQQHFWV�PRUH�WKDQ���������PHPEHUV��IRVWHULQJ�FURVV�JRYHUQPHQW�FROODERUDWLRQ��VROYLQJ�FRPPRQ�SUREOHPV�DQG�DGYDQFLQJ�JRYHUQPHQW�FDUHHUV��*RY-/RRS�LV�KHDGTXDUWHUHG�LQ�:DVKLQJWRQ��'�&���ZLWK�D�WHDP�RI�GHGLFDW-HG�SURIHVVLRQDOV�ZKR�VKDUH�D�FRPPLWPHQW�WR�FRQQHFW�DQG�LPSURYH�government. )RU�PRUH�LQIRUPDWLRQ�DERXW�WKLV�UHSRUW��SOHDVH�UHDFK�RXW�WR� LQIR#JRYORRS�FRP. ZZZ�JRYORRS�FRP | @GovLoop
ThDQN�\RX�WR�ΖQIRUPDWLRQ�%XLOGHUV��1HW$SS��2UDFOH��3XUH�6WRUDJH��7HJLOH��7HUDGDWD�DQG�9L21�IRU�WKHLU�VXSSRUW�RI�WKLV�YDOXDEOH�UH�VRXUFH�IRU�SXEOLF�VHFWRU�SURIHVVLRQDOV�
Author: Nicole Blake Johnson, Technology Writer
Designers: -H�5LEHLUD��&UHDWLYH�0DQDJHU Tommy Bowen, Graphic Designer .DLWO\Q�%DNHU��-XQLRU�'HVLJQHU0DUWLQ�1HUD��'HVLJQ�)HOORZ
38 | A GOVLOOP GUIDE
1152 15th St NW, Suite 800
Washington, DC 20005
Phone: (202) 407-7421 | Fax: (202) 407-7501
www.govloop.com
@GovLoop