Dashboard for Statistical Visualization of Client Service Events Alex Zolot (Zolotovitski), PhD [email protected] .

Dashboard for Statistical Visualization of Client Service Events

Alex Zolot (Zolotovitski), [email protected] www.StatVis.com

www.zolot.us www.StatVis.com 2

2


The raw data:Creation – Received –First_Contact_SW – First_Contact_HW – Request – Pending – Closed

TSC Events: Creation – Received – First_Contact_SW –

First_Contact_HW – Request – Comp_Pending – Closed

Types: HW Platforms, Products, Geo, Owner, …


Case DateTime Event Type Time PrEvCase1 T1 Ev1 Type1 0 0Case1 T2 Ev3 Type1 T2-T1 E1

Case1 T3 Ev7 Type1 T3-T2 E3

Case2 T4 Ev3 Type1 0 0……

Sorted data


OLAP dimensions.



Consists of three key elements:

• Run chart for longitudinal (in time) representation of data • Matrix bar chart for types of cases• Flower bed chart for time between events representation.

The flower bed and matrix bar charts can be plotted for any specific interval of time: week, month, or year.

All features of these three charts are clickable, switching is easy between charts, and drilling down can be done to detailed lists of individual cases. These interactive capabilities facilitate identification of the most business important cases.


6

Longitudinal Analysis: Run Chart

TTR = Time-To-Resolve = DateTime(CaseClosed) – DateTime(CaseCreated)


7

Cross-Sectional Analysis –Time Between Events:

The raw TSC data:Creation – Received –First_Contact_SW – First_Contact_HW – Request – Pending – Closed

Events: Creation – Received – First_Contact_SW – First_Contact_HW – Request – Comp_Pending – Closed

Types: HW Platforms, Products, Geo, Owner, …


8



9



10


• Traditional way to represent 2 (pivot) tables: Freq and MTBE (Event x PrevEvent )

Freq

0

20000

40000

60000

80000

100000

R S H Q P C W

Event

Freq

CREATION

RECEIVED

FIRST_CONTACT_SW

FIRST_CONTACT_HW

REQUEST

COMP_PEND

CLOSED

WAREHOUSE_UPDATE

TTR

0

100

200

300

400

500

600

700

800

900

R S H Q P C W

EventTT

R, d

ays

CREATION

RECEIVED

FIRST_CONTACT_SW

FIRST_CONTACT_HW

REQUEST

COMP_PEND

CLOSED

WAREHOUSE_UPDATE

MTBEMTBE

MTB

E


11

Cross-Sectional Analysis –Time Between Events: Variants of possible designs for 2 parameters (Fr and T)

• Area of Circle (“house”) is proportional to frequency of the event.• Width of Petal (arrow, bar) is proportional to frequency of transitions between events.• Length of Petal is proportional to mean time between events.• Color of Petal is the same as color of “To” Circle (“house”)


12

Cross-Sectional Analysis –Time Between Events: Flower Bed Chart

Data visualization – variants of designPetal direction points to next event, with matching petal-circle color. Circle size ~ frequency. Length of petal ~ time and width of petal ~ frequency.


13

Cross-Sectional Analysis – Tower Charts

1 1

3 3

7

7

3

3

1

1

Figure 5. Empirical CDF (left) and sorted sequence (right) of TBE = {1, 1, 3, 3, 7}.

f(i) = Q( i / N) = ECDF-1(i / N)

f(i) = Q( i / N) = ECDF-1(i / N)

. 3D Towers

3D Towers


14

Cross-Sectional Analysis –Time Between Events: Flower Bed Chart with Tower Petals


15

Cross-Sectional Analysis – Evolution of Elements


Cross-Sectional Analysis by Types of Cases:

Traditional way To represent 2 (pivot) tables: Freq and TTR (HW Platform) x (Product)


Cross-Sectional Analysis by Types of Cases: Matrix Bar Chart

Width of cuboid is proportional to time and height is proportional to square root of frequency.





Alex Zolot (Zolotovitski), [email protected] www.StatVis.com

Dashboard for Statistical Visualization of Client Service Events Alex Zolot (Zolotovitski), PhD [email protected] .

Documents

crosssectional analysis

house slide

closed events

events representation

sorted data slide

closed tsc events

longitudinal analysis

specific interval of