Difference between revisions of "ISSS608 2017-18 T3 Assign Aakanksha Kumari Q3"

From Visual Analytics and Applications
Jump to navigation Jump to search
 
(12 intermediate revisions by the same user not shown)
Line 2: Line 2:
  
 
<div style="background:#DC143C; border:#DC143C; padding-left:15px; text-align:center;">  
 
<div style="background:#DC143C; border:#DC143C; padding-left:15px; text-align:center;">  
<font size = 5; color="#FFFFFF"><span style="font-family:Century Gothic;">MC Challenge 3</span></font>  
+
<font size = 5; color="#FFFFFF"><span style="font-family:Century Gothic;">Unraveling the Secrets of Kasios : VAST Mini Challenge 3</span></font>  
 
</div>
 
</div>
 
<!--MAIN HEADER -->  
 
<!--MAIN HEADER -->  
Line 29: Line 29:
 
;  
 
;  
 
[[ISSS608 2017-18 T3 Assign Aakanksha Kumari_Q4| <font color="#FFFFFF">Question 4</font>]]  
 
[[ISSS608 2017-18 T3 Assign Aakanksha Kumari_Q4| <font color="#FFFFFF">Question 4</font>]]  
 
| style="font-family:Century Gothic; font-size:100%; solid #DC143C; background:#DC143C; text-align:center;" width="14.3%" | 
 
;
 
[[ISSS608 2017-18 T3 Assign Aakanksha Kumari_conclusion| <font color="#FFFFFF">Conclusion</font>]]
 
  
 
| style="font-family:Century Gothic; font-size:100%; solid #DC143C; background:#DC143C; text-align:center;" width="14.3%" |   
 
| style="font-family:Century Gothic; font-size:100%; solid #DC143C; background:#DC143C; text-align:center;" width="14.3%" |   
Line 101: Line 97:
 
=== Clusters Discussed Individually ===
 
=== Clusters Discussed Individually ===
 
==== Cluster 1 (Red) ====
 
==== Cluster 1 (Red) ====
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Main Channel of Interaction''': Purchase  <br>
 +
'''Key Players''': Gail Feindt, Richard Fox and Meryl Pastuch
 +
</big> </div>
 +
|}
 +
 +
[[File:Cluster0-Red.png|350px|frameless|center]]
 +
 +
==== Cluster 2 (Dark Green) (Small cluster, connected through call)  ====
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Main Channel of Interaction''': Call  <br>
 +
'''Key Players''': Lizbeth Jindra, Madeline Nindorf
 +
</big> </div>
 +
|}
 +
[[File:Cluster1-DarkGreen.png|450px|frameless|center]]
 +
 +
==== Cluster 3 (Pink)  ====
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Main Channel of Interaction''': Email, Call  <br>
 +
'''Key Players''': Lindsy Henion, Sheilah Stachniw, Kerstin Belveal
 +
</big> </div>
 +
|}
 +
 +
[[File:Cluster2-Pink.png|700px|frameless|center]]
 +
 +
==== Cluster 4 (Light Green)  ====
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Main Channel of Interaction''': Email, Call  <br>
 +
'''Key Players''': Sherrell Biebel, Tobi Gatlin, Julie Tierno
 +
</big> </div>
 +
|}
 +
 +
[[File:Cluster3-LightGreen.png|600px|frameless|center]]
 +
 +
==== Cluster 5 (Blue)  ====
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Main Channel of Interaction''': Email  <br>
 +
'''Key Players''': Glen Grant
 +
</big> </div>
 +
|}
 +
 +
[[File:Cluster4-Blue.png|600px|frameless|center]]
 +
 +
=== Group composition over different activities ===
 +
==== Calls ====
 +
 +
<div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
Large number of call happens intra cluster. But a few calls do happen inter clusters. Most of calls which are across clusters, are to Purchase cluster. </big> </div>
 +
 +
[[File:Calls1.png|1000px|frameless|center]]
 +
 +
==== Emails ====
 +
<div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
Primary method of communication within each cluster. Gail Feindt is not connected to any node thorugh email. Major recipents of email outside cluster is Richar Fox and Meryl Pastuch.  </big> </div>
 +
 +
[[File:Emails1.png|1000px|frameless|center]]
 +
 +
==== Meetings ====
 +
<div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
'''Meetings shows the most interesting pattern.'''
 +
*Meetings mostly happen inter cluster.
 +
*Blue Cluster, did not participate in any meetings. Which clears them of any suspicion.
 +
*Meetings primarly connect Red cluster, yet again.
 +
*Gail Feindt is again not connected to any employee through meetings.</big> </div>
 +
 +
[[File:Meetings1.png|1000px|frameless|center]]
 +
 +
 +
==== Purchases ====
 +
 +
[[File:Purchase1.png|1000px|frameless|center]]
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
 +
'''Purchases (combined with meetings) hints towards source of all suspicion.'''
 +
*Gail Feind in the center of it all.
 +
*No purchases intra-cluster.
 +
*Key players iniating purchases are Richar Fox, Meryl Pastuch, Tobi Gatlin, Sheilah Stachniw, Lizbeth Jindra
 +
*Each cluster other than blue is making purhcases with Gail Feind. (As previously mentioned employees in blue cluster were already clear of suspicion)
 +
*Tobi Gatlin, Sheilah Stachniw, Lizabeth Jindra did not attend any suspicious meetings with Richard Fox or Meryl Pastuch. This probably mean that these employees are only ordering purchases from Gail Feind from orders passed by other players who actually attend the meetings. This implies that these employees are only means of making suspicious purchases, the strings are pulled from behind by other key employees who attend the meetings and are in close contact with Richard Fox and Meryl Pastuch.
 +
</big> </div>
 +
|}
 +
 +
=== Group’s Interactions Change Over Time ===
 +
<gallery mode=packed widths="800px" heights="800px">
 +
File:2015.png|2015
 +
File:2016.png|2016
 +
File:2017.png|2017
 +
</gallery>
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
 +
'''Purchases (combined with meetings) hints towards source of all suspicion.'''
 +
*No suspicious purhcases takes place in 2015.
 +
*Only one suspicious purhcases takes place in 2016 (November 21), with all other happeing in 2017. This means all the suspicious were mostly carries during end of 2016 and 2017..
 +
*2017 also marks a decrease in communication with other employees, many of employees were not contacted in 2017. This might mean either these employees left the organisation (refer Q1) or they were not kept in chain of suspicious communication any longer.
 +
*The blue cluster is slowly phased out over the years.
 +
 +
</big> </div>
 +
|}
 +
 +
=== Conclusion ===
 +
 +
{| class="wikitable"
 +
|-
 +
| <div style="font-family:Palatino Linotype; border-radius: 1px "> <big>
 +
 +
*We began with 26 suspicious employees, but now after analysiing their communication pattern, we can narrow down to fewer employees.
 +
*Gail Feindt is destination of each purchase transaction.
 +
*Richard Fox and Meryl Pastuch are very important employees in this investigation. These two employees hold lot of influence.
 +
*Most of suspicious purchases happended in year 2017 and end of 2016.
 +
 +
</big> </div>
 +
|}
 +
 +
=== Narrowed down suspicious employees (12) ===
 +
 +
'''''<big>'''Gail Feindt, Richard Fox, Meryl Pastuch, Lizbeth Jindra, Madeline Nindorf, Lindsy Henion, Sheilah Stachniw, Kerstin Belveal, Sherrell Biebel, Tobi Gatlin, Julie Tierno, Glen Grant'''</big>'''
 +
''

Latest revision as of 13:09, 8 July 2018

Classified-stamp.png

Unraveling the Secrets of Kasios : VAST Mini Challenge 3

Overview

Data Preparation

Question 1

Question 2

Question 3

Question 4

Dropbox

 

Using the combined group of suspected bad actors, you created in question 2, show the interactions within the group over time.

  • Characterize the group’s organizational structure and show a full picture of communications within the group.
  • Does the group composition change during the course of their activities?
  • How do the group’s interactions change over time?

The Story So Far

The Kasios International over the course of two and half years have maintained nearly same communicationa volumne and purchase pattern, which implies that the organization have not grown or shrunk significantly. But on closer analysis we found out with each year, the new employees were being hired, and there is constant shuffling of human resource. Kasios have grow significant change in personnels, with numnber of same employees each month decreasing.

Intially we began with certain suspicious transactions (purchaes and communication) as provided by our informer. Later as we continued our analysis, othe key players, closely associated to employees making suspicious communication were identified. Using network analysis and eigenvector centrality (filter value > 0.4), nodes connected to influential nodes (in this case, suspicious employees) more than other non-influential nodes(non suspicious employees), can be identified. So by following the above startegy following are list of old and new suspicious players.

Original Suspicious Employess

Augusta Sharp, Tobi Gatlin ,Patrick Lane, Melita Scarpaci, Rosalia Larroque, Glen Grant ,Meryl Pastuch, Dylan Ballard, May Burton, Kerstin Belveal, Alex Hall, Refugio Orrantia, Jose Ringwald, Julie Tierno, Richard Fox, Lizbeth Jindra, Ramiro Gault, Sara Ballard, Jenice Savaria, Lindsy Henion

New Suspicious Employees
Ferne Hards, Martha Harris, Sherrell Biebel, Gail Feindt, Sheilah Stachniw, Madeline Nindorf

So now our focus is on these 26 key players, as discussed in Q2

Main Tasks in this section
Using the combined group of suspected bad actors you created in question 2, show the interactions within the group over time.

  • Characterize the group’s organizational structure and show a full picture of communications within the group.
  • Does the group composition change during the course of their activities?
  • How do the group’s interactions change over time?

The Combined Communication Picture

Suspicious communication within group-alt.png

Using modularity algoithm, and assigning distinct class color, the graph is clustered into 5 classes. This helps in better visualization, and also identifying key players in each cluster. All the clusters are discusses individually later.

What can we learn

  • Purchases are few in between. Key players involved are Gail Feindt, Richard Fox and Meryl Pastuch.
  • Organization communication structure is hierarchial. Which means only few key players from each cluster communicates with key players from other cluster.
  • Less meetings, mainly happenig between inter cluster groups.

Clusters Discussed Individually

Cluster 1 (Red)

Main Channel of Interaction: Purchase
Key Players: Gail Feindt, Richard Fox and Meryl Pastuch

Cluster0-Red.png

Cluster 2 (Dark Green) (Small cluster, connected through call)

Main Channel of Interaction: Call
Key Players: Lizbeth Jindra, Madeline Nindorf

Cluster1-DarkGreen.png

Cluster 3 (Pink)

Main Channel of Interaction: Email, Call
Key Players: Lindsy Henion, Sheilah Stachniw, Kerstin Belveal

Cluster2-Pink.png

Cluster 4 (Light Green)

Main Channel of Interaction: Email, Call
Key Players: Sherrell Biebel, Tobi Gatlin, Julie Tierno

Cluster3-LightGreen.png

Cluster 5 (Blue)

Main Channel of Interaction: Email
Key Players: Glen Grant

Cluster4-Blue.png

Group composition over different activities

Calls

Large number of call happens intra cluster. But a few calls do happen inter clusters. Most of calls which are across clusters, are to Purchase cluster.
Calls1.png

Emails

Primary method of communication within each cluster. Gail Feindt is not connected to any node thorugh email. Major recipents of email outside cluster is Richar Fox and Meryl Pastuch.
Emails1.png

Meetings

Meetings shows the most interesting pattern.

  • Meetings mostly happen inter cluster.
  • Blue Cluster, did not participate in any meetings. Which clears them of any suspicion.
  • Meetings primarly connect Red cluster, yet again.
  • Gail Feindt is again not connected to any employee through meetings.
Meetings1.png


Purchases

Purchase1.png

Purchases (combined with meetings) hints towards source of all suspicion.

  • Gail Feind in the center of it all.
  • No purchases intra-cluster.
  • Key players iniating purchases are Richar Fox, Meryl Pastuch, Tobi Gatlin, Sheilah Stachniw, Lizbeth Jindra
  • Each cluster other than blue is making purhcases with Gail Feind. (As previously mentioned employees in blue cluster were already clear of suspicion)
  • Tobi Gatlin, Sheilah Stachniw, Lizabeth Jindra did not attend any suspicious meetings with Richard Fox or Meryl Pastuch. This probably mean that these employees are only ordering purchases from Gail Feind from orders passed by other players who actually attend the meetings. This implies that these employees are only means of making suspicious purchases, the strings are pulled from behind by other key employees who attend the meetings and are in close contact with Richard Fox and Meryl Pastuch.

Group’s Interactions Change Over Time

Purchases (combined with meetings) hints towards source of all suspicion.

  • No suspicious purhcases takes place in 2015.
  • Only one suspicious purhcases takes place in 2016 (November 21), with all other happeing in 2017. This means all the suspicious were mostly carries during end of 2016 and 2017..
  • 2017 also marks a decrease in communication with other employees, many of employees were not contacted in 2017. This might mean either these employees left the organisation (refer Q1) or they were not kept in chain of suspicious communication any longer.
  • The blue cluster is slowly phased out over the years.

Conclusion

  • We began with 26 suspicious employees, but now after analysiing their communication pattern, we can narrow down to fewer employees.
  • Gail Feindt is destination of each purchase transaction.
  • Richard Fox and Meryl Pastuch are very important employees in this investigation. These two employees hold lot of influence.
  • Most of suspicious purchases happended in year 2017 and end of 2016.

Narrowed down suspicious employees (12)

Gail Feindt, Richard Fox, Meryl Pastuch, Lizbeth Jindra, Madeline Nindorf, Lindsy Henion, Sheilah Stachniw, Kerstin Belveal, Sherrell Biebel, Tobi Gatlin, Julie Tierno, Glen Grant