1 ANA02 Visual story telling The key to unlocking the data gold mine
1
ANA02 Visual story telling
The key to unlocking the data gold mine
2
Robert van den Breemen
@rvdbreemen
https://nl.linkedin.com/in/rvdbreemen
Femke Goedhart
@femkegoedhart
https://www.linkedin.com/in/femkegoedhart
3
The saga continues!
4
Time sequenceMultiple Dimensions / AxisesClutteredBusy diagram
But what does it mean?
Report design from 2011 onward
In the beginning… Excel
5
To measure = to know!
Tripling of documents on
traditional file shares since 2010
Suspecting of high number
of duplications…
Digital Enterprise
Dutch Tax and Customs
Administration
Lack of insight!
Exponential
Growth?
Users wonder
where to store
files / documents
WHY?
6
Different end user devices
unlimited
unlimited
unlimited
unlimited
unlimited unlimitedunlimited
unlimited
unlimited
unlimitedunlimited
unlimited
Always, unlimited,
everywhere available
Data is everywhere
is stored on multiple locations
is unfindable
is never deleted
Data grows and grows…..
Data is not accessible on mobile
Data is lost
Data is created multiple times
Different locations (anywhere)
Data growth the Enterprise reality?!
The need for…
WHY?
7
So where do we store our files?
FileserversEOL, EOC
Home directory,
profile
Shared folders
Application data
1,2 billion files450 servers2.000 TB storage
Classical Windows
Fileshares
> 1,5 million files2 serversSome 2 TB storage
IBM Connections
(incl. Libraries)
40.000 users30 servers7.000 team mailboxes500 TB storage
Mailboxes, Team
mailboxes and mail
archive
Xxxx millionsdocumentsSome serversXxxx TB storage
Document
management (ECM)
Scope of the current project
30.000+ laptops/desktopsUnknown files (±100.000)Unknown amnount of storage
Local files on laptops /
desktops
WHAT?
8
+
9
Collections
of the metadata FileserversEOL, EOC
Home directory,
profiel
Samenwerkings
Gebieden
Applicatiedata
Process
Exploration of the
Collected meta-data
Insights
Fact-based decisions
and strategy’s
1. Technical measures (cost saving potential, de-duplication, application)
2. Information for lines of business (accountable)
3. Change the user behavior
(adoption strategy)
To measure = to know! To see is to believe!
To understand,
means you can do something!
To discover what you did not know…
HOW?
10
The Perfect Storm
2017
GDPR
Dataleaksin the news(breaking news)
Court orders todelete images of ”Licence plates"
Cleaning up the Network Shares
FRA
Retention
laws
Idenitifyingshadow IT
Applicationrationalisation
11
To see is to know!
Are we ready for GDPR?
Increasing need
for Insight!
Deeper insight needed!
Digital Enterprise
Dutch Taxs and Customs
Administration
Can we delete the licence plates?
NOW
12
Collect dataVisual
ExplorationRefine
Create
Report
Data
Graph
13
Initial findingsTrying different thingsDiscover
Relative meaning
What is the relation here?
Results from Visual Discovery Phase
Report design from 2016
14
My inspirational moment: Hans Rosling
Source: TED talksWebsite: https://www.gapminder.org/
15
“It hit me...”
16
Collect dataVisual
ExplorationRefine
Create
Story
Visual
Story
From Reporting
To Storytelling
17
Story building
Domain knowledge
Processing skills
Story
Telling
Domain knowledge
Design
skills
Data Engineering
skills
Exploration
“I know how tovisualize”
“I know how towork with data”
“I know what theoutcome shoud be”
+
18
Building the visual story
19
Start at the beginning...
WHO is your audience WHAT is your call to actionWHAT is your medium
HOW are you visualizing it
20
• Invites to explore
• Offers insight directions
• Promotes discussion and exploration
• Invites to learn
• Offers directed insights
• Often leads to a conclusion or action
Exploratory vs Explanatory
Leads to better
decision making
21
Exploratory Example
22
Explanatory examples
23
A few pointers for design….
24
High quantative perception
Low quantative perception
Which type of graph to use
25
Preattentive processing
Source: Design for Information: An Introduction to the Histories, Theories, and Best Practices Behind Effective Information VisualizationsIsabel Meirelles
26
The usage of colors as preattentiveattributes
27
• Titles/subtitles
• Labels
• Callouts
• Legends
• As a graph
Text is your friend!
28
Hard to quantivy dimensions (3D!)
No preattentive attributes (color use?)
What am I looking at…(no text)?
What to avoid…
Image source: https://stackoverflow.com/questions/22369224/3d-pie-chart-in-highcharts-javascript
29
An exampleACME INC wants to know if they can justify a Enterprise SocialNetwork based on the files storage in their current systems
30
31
A lot of data… But no story
32
Lets get started
• Sorting
• Color use
• Remove obsolete elements
• Grouping (“Other”)
• Type of graph
33
Let’s refine it
• Change orientation & collate
• Remove obsolete elements (Legend)
• Smart labeling
• Take out grid lines
• Take out unnecessary column headers
• Add tooltips
• Take out legend
34
Let’s bring it all together...
• Add titles, explanation and “Call to action”
• Take out everything that isn’t necessary
• Zig-zag structure
35
Data Storytelling:
data in context telling the story
Data Visualization:
mapping data on a graph
36
37
With insight… … time for action!
38
Leasons learned…
• The data must find the data and the relevance must find you
• Visualization is easy and key to excite users, however visual storytelling is hard
• Iterate, iterate, iterate, keep an open mind!
39
Our next steps
Back to the original data goldmine idea:
• Bring more meta-data into the mix than just file meta-data, for example mail-interactions, phone-records, chat-interaction, ESN data, USB use, applicatoin usage data, etc.
• Start to mine the file-content with content analytics. To build a GDPR scanner and classify documents in business relevant classes to enable archiving.
41
Questions?
Robert van den Breemen
@rvdbreemen
https://nl.linkedin.com/in/rvdbreemen
Femke Goedhart
@femkegoedhart
https://www.linkedin.com/in/femkegoedhart
42
The END