Entry Name: "VASE UCALGARY-Challenge1"
VAST
Challenge 2014
Mini-Challenge 1
Team Members:
Dr. Craig Anslow,
University of Calgary, craig.anslow@ucalgary.ca
Dr. Frank Maurer, University of Calgary, frank.maurer@ucalgary.ca
Dr. Mario Costa Sousa's, University Of Calgary, mario@cpsc.ucalgary.ca
Dr. Faramarz
F. Samavati, University of Calgary, samavati@cpsc.ucalgary.ca
Student Team: YES
Rahul
Kamal Bhaskar, University of Calgary rbhaskar@ucalgary.ca PRIMARY
Julia
Parades, University of Calgary, jparedes1006@gmail.com
Zahra
Shakeri, University of Calgary, lloi.shakeri@gmail.com
Zahra
Sahaf, University of Calgary, sce2020sahaf@gmail.com
Haleh Alemasoom,
University of Calgary, h.alemasoom@gmail.com
Analytic Tools Used:
Excel
D3.js
Highchart.js
Vast 2014 (Tool
developed by our team)
Approximately how many hours were spent
working on this submission in total?
We spend total: 172 hours 15 minutes
Discussion: 1*6
hours = 6 hours
Coding: 156 hours
15 minutes
Final answer
Write-ups: 10 hours
May we post your submission in the Visual
Analytics Benchmark Repository after VAST Challenge 2014 is complete?
YES
Video:
Google Drive link: https://docs.google.com/file/d/0B4qcf3SpiLhcaVB6c0lLeTdzZnc/edit
YouTube Link: https://www.youtube.com/watch?v=c63DmwfkMmM&feature=youtu.be
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Questions
MC1.1
– Provide a visual
representation of the structure of the Protectors of Kronos network, with
supporting evidence.
a. Who are the leaders?
b. Who is part of the extended network?
c. How has the group structure and organization
changed over time?
d. Where are the potential connections between the
POK and GAStech?
Provide novel visualizations
appropriate for communicating key information to the busy leaders of the
investigation. Please limit your response to no more than eight images and 500
words.
Answer
Figure 1: Infographic showing organizational change over
time and people involved in POK
a.
Who are the leaders?
To solve this problem we used both manual (i.e. by reading
historical documents) and automatic approach (By searching keyword leader in
our system).
|
Name of Leader |
Time Period |
Source of data |
|
Henk Bodrogi |
1997-2001 |
Historical documents: 5 year report, 10 year historical
document |
|
Elian Karel |
2002-2004 |
Historical documents: 5 year report, 10 year historical
document |
|
Elian Karel |
2005-2009 |
Historical documents: 5 year report, 10 year
historical document Article document number: 454 |
|
Silvia Marek |
2009-Present |
Article document number: 454 |
Table 1.1.1: Showing leader and their tenure
We visualize this graph by
showing the novel infographic visualization. It shows leader with animated
person and time period when they are elected.
b.
Who
is the part of extend network?
In order to answer this question we first queried the system from
(1992.11.12) to (2014.01.?) and generated ‘Network Graph’ based on articles for
this time period. As shown in Figure 1.1.1, this graph shows two types of
relationships:
(1) Relationship between employees and their organizations (POK, GASTech)
(2) Relationship between different employees; two employees are
connected if they have been appeared in the same article (figure 1.1.1).
Figure 1.1.1: Article Network Graph for POK and GASTech
By clicking on each node of this graph, we can see information
about employee name, the department in which the employee is working and
his/her role (see Figure 1.1.2).
Figure 1.1.2: Network Graph showing node value on mouse hover
We used this information to find employees who are related to both
organizations and considered them as a member of Extended Network. As
you see in Figure 1.1.3, we searched
all of employees’ names that appeared in common section and skimmed their
related articles.
Figure 1.1.3: Query selection for the employees in article
At this step we did not find any employee as a member of Extended
Network. Isia Vann was the only person in our common
section of our network graph that his name was not found in articles. As mentioned
before, we excluded ‘5 year report’ and ‘10 year historical document’ from our
database, but we used it for network graph. After skimming this profile document, we found that: “Isia Vann is a POK
member and GasTech employee and is a part of Extended
Network”.
c.
How has the group structure and organization changed
over time?
Answer:
Before Juliana’s death, POK was
composed by the founding members and their families (figure 1). More people
joined the group after Elian Karel changed their agenda to include protests
against the corrupt government. POK protests became more violent with the
increasing number of members after the Tiskele River
caught fire in 2005. In 2005, all members of the pacific group Save Our Wildlands joined POK. These members included people like
Silvia Marek and Lucio Jakab. After Elian’s death,
Marek took charge of POK.
d.
Where are the potential connections between the
POK and GAStech?
Answer:
To answer this question we have
created a network graph in which there were two types of nodes organization
(i.e. GASTech and POK) and people (i.e. GASTech employee and POK founders). In this GASTech employee’s nodes are connected to the GASTech organization node and POK founder’s nodes are
connected to POK organization node. Now for finding potential connection we
prepared a graph based on the connection between people node and organization.
This connection was based on the parsing of the articles if the two name (i.e.
people or organization) appeared in the same article we created a link between
it. This was done in assuming that if they are mentioned in the article then
there might be some connection between them. This can be verified by reading
that article. We also connected nodes on the basis of last name. Below is the
graph (figure 1.1.2) which shows potential connection. Nodes which are present
between GASTech and POK might be potential
connection. Kare Orilla, Anda Ribera, Isia Vann, Edvard Vann, Ada Campo-Corrente,
Elian Karel, Jeroen Karel.
Figure 1.1.2: Network Graph
We used the last names of founding members of POK to know if their
families worked at GasTech. The following employees
search last names with one of the founding members; however, only Edvard Vann has been questioned by the police (Table:
1.1.2.):
|
POK possible relative |
Employee |
Type |
Title |
Source Of Data |
|
Carmine Osvaldo |
Hennie Osvaldo |
Security |
Perimeter Control |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Valentine Mies |
Henk Mies |
Facilities |
Truck Driver |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Valentine Mies |
Minke Mies |
Security |
Perimeter Control |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Valentine Mies |
Ruscella Mies |
Administration |
Assistant Manager |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Henk Bodrogi
|
Loreto Bodrogi |
Security |
Site Control |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Juliana Vann and Mandor Vann |
Edvard Vann |
Security |
Perimeter Control |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
|
Juliana Vann and Mandor Vann |
Isia Vann |
Security |
Perimeter Control |
Historical documents: 5 year report, 10 year
historical document and Employee Record |
Table 1.2.2: Possible connection between
POK member and GASTech Employee
MC1.2
– Describe the
events of January 20-21, 2014. What is the timeline of events? Please limit
your response to no more than ten images and 500 words.
To find answer to this question we
designed a system (figure 1.2.1)
that performed natural language processing on the data set. First we filtered
documents by performing a general query using “Query Section” to search a word within a particular date range for
the documents (i.e. articles) which we are concerned with. Once articles has
been filtered Viz1,Viz2, Viz3 and document section (Viz1: Is a bar chart showing count of articles date wise, Viz2: Is a word cloud showing most used
word in articles; Viz3: Is also a
word cloud showing most used word in categorized way, for example, name or
person, organization, money etc.; Document
section: shows all articles in the selected date range) has appeared, we
start checking in the “Classified WordList” section
to see the most used word or suspicious word in the document. If we find
something relevant to that word we select that word then we have list of all
articles (Viz 4) which has selected word. In order
to view that word which we have selected is relevant we select article from the
list and check highlighted document in “Corresponding
Selected Article” that this word is able to answer the question or not in
this section we have highlighted the text and provided detail about the color
in legend.
Figure 1.2.1: System snapshot
For this question we have searched
for the word “event” between
01/01/2014 to 01/31/2014.
Then looked at most
frequent word appeared in articles related to the dates in classified wordlist.
Figure 1.2.2: Date
Classified word for event keyword between 01/01/2014 to 01/31/2014
In word cloud (figure 1.2.2) all dates seems to be related to
January 2014 except “twentieth year”. So we checked the articles in which it
has appeared.
There are two articles
where this dates appeared. Article no 62 and 614
Figure 1.2.3: Article 62
On
reading both the articles (figure 1.2.2)
we can conclude that event was organized to celebrate twentieth year of the
cooperation between GASTech and Abila Government.
Then
we also looked at the classified word section related to time (figure 1.2.4). To find the timings of the
event. (Note: Color has no significance in the word cloud only size in
significant. Bigger the font more it has occurred in document)
Figure 1.2.4: Time
Classified word for event keyword between 01/01/2014 to 01/31/2014
In this cloud “this
morning” is the most used phrase so we selected it and got list of article (figure
1.2.5) which has this word.
Figure 1.2.5: List of the
articles for “This Morning” time.
If user wants to verify
that they have not missed any document then can look for documents in the list
of all article section (figure 1.2.6)
where we can check for relevant heading.
Figure 1.2.6: List of all
articles within the selected time period
So here article number
140 seems to be relevant to event.
Article (figure 1.2.7) it clearly states the timeline
of the event i.e. Morning event and the reception with government. But event
was stopped due to the fire alarm.
Figure 1.2.7: Article 140
Now
for finding event related to kidnapping. I performed following task:
1. Searched keyword “kidnap”. For date 01/01/2014 to
01/31/2014. As we were concerned about time so looked for blog articles as they
contains time element. In reading some article we figured out that conference
was conducted to report the progress on the case.
2. Looked for articles
related to “Conference” as we can
extract timing details of conference held by police as well as GASTech. We selected “Abila police” from classified
wordlist and organization section.
3. While reading blog
articles we assume the at the beginning of blog is
time in 24 hour time format. Before 12:45 on 01/20/2014 cops secured the GASTech headquarter perimeter (Figure 1.2.8).
Figure 1.2.8: Article 356
4. Searched keyword plane:
Information derived two plane departed one at 12:30 p.m. and other at 2:30 p.m.
on 01/20/2014 (Figure 1.2.9).
Figure1.2.9: Article 718
5. There was police
conference on 01/20/2014 before 19:47 (figure 1.2.10)
Figure 1.2.10: Article
139
6. Blog post shows that
Abila police conference at 9:00 a.m. on 01/21/2014. For kidnapping confirmation
and the number of kidnapped person.
(Figure 1.2.11)
Figure 1.2.11: Article
276
7. Word searched “investigate” selected blog articles. In
this it stated that GAStech International news
conference 10:00 am 01/21/2014
Figure 1.2.12: Article
250
Summary of above analysis on
overall timeline of evnets on 20-21 January is
|
|
Date |
Time |
Event |
|
1 |
01/20/2014 |
In morning |
Company has agenda of annual meeting
followed by reception of government of Kronos |
|
2 |
01/20/2014 |
10:00 am |
Annual meeting was closed at 10:00
due to fire alert |
|
3 |
01/20/2014 |
Before 12:45 p.m. |
cops secured the GASTech headquarter perimeter |
|
4 |
01/20/2014 |
12:30 |
Plane departed to unknown location.
Passenger looks worried. |
|
5 |
01/20/2014 |
2:30 p.m. or 14:30 |
Plane departed to Rome Italy.
Passenger looks happy as celebrating some thing |
|
6 |
01/20/2014 |
before 19:47 |
Police conference that people are missing
but not sure whether they are kidnapped or not |
|
7 |
01/21/2014 |
9:00 a.m. |
For kidnapping confirmation and the
number of kidnapped person |
|
8 |
01/21/2014 |
10:00 a.m. |
GAStech International news
conference |
|
9 |
|
|
|
|
10 |
|
|
|
|
11 |
|
|
|
MC1.3 – Identify at
least two possible explanations why the GAStech
employees may be missing. What evidence do you have to support each of these
explanations? Please limit your response to no more than three additional
images and 200 words.
Answer
1. We searched the keyword “missing”
2. In the classified word
section looked in money section to find whether any illegal activity has been
performed for money and can be linked to the missing people.
Figure 1.3.1: Money
classified word cloud for “missing” for 01/20/2014 to 01/31/2014
3. Application showed up “20
Million” in word cloud. On selecting this value. In article no 167 (figure 167)
it’s mentioned that POK claimed responsibility of kidnapping and demand 20
million from the company.
Figure 1.3.2: Article 167
So missing people might
have been kidnapped by the POK. This may be the first possible reason.
4.
Now we selected word (i.e. “Abila Police”) from “Classified WordList” from organization section. To check opinion of
police on the kidnapping. In this we verified multiple article and found
article 250 (figure 1.3.4) valuable
it stated that number of kidnapping has been revised from 14 to 10. So
according to this article 4 people who were reported missing earlier were
found.
Figure 1.3.4: Article 250
5. Selected conference from
the “Overview Wordlist” and then
selected article with headline “GAStech Sanjorge escaped from the Kidnapping at Gastech
HQ”. As you can see in article 167 it’s stated that 5 executives along with
C.E.O Sten Sanjorge Jr. was
missing. So exploring his escape can give us second reason behind missing
assumption of these people.
Figure 1.3.5: Article 689
Figure 1.3.6: Article 344
In article (figure 1.3.5
and 1.3.6) it is stated that Sten Sanjorge Jr. escaped
kidnapping as he was in transit when the kidnapping occurred. So from these articles
we can assume that people who were assumed missing and found later may be in
transit during the kidnapping.
Summary:
Total
people assumed to be missing was 14 on 01/20/2014. After investigation police
reviewed their count from 14 to 10. According to analysis possible reasons are
as follows:
·
1st possible reason: 10
missing people are kidnapped by POK.
·
2nd possible reason: 4 who
were assumed missing people were on transit from GASTech
to Capitol building.