Pennsylvania State University Senseplace and GeoVizToolkit
VAST 2011 Challenge
VAST 2011 Challenge
Mini-Challenge 3 - Investigation into Terrorist Activity
Authors and
Affiliations (in alphabetical order):
Students:
Ying Chen2 yxc242@ist.psu.edu
Nicklaus
Giacobe2 nxg13@ist.psu.edu
Anuj Jaiswal2 ajaiswal@psu.edu
Wei Luo1 wul132@psu.edu
Alexander Savelyev1 savelyev@psu.edu
Vitalie Victorov2
vwv5007@psu.edu
Sen Xu1
sux100@psu.edu
Faculty &
Staff:
Justine Blanford1 jib18@psu.edu
Frank Hardisty1
hardisty@psu.edu
Alan MacEachren1 macheachren@psu.edu
Prasenjit Mitra2
pmitra@ist.psu.edu
Scott Pezanoski1 spezanowski@psu.edu
Anthony Robinson1 arobinson@psu.edu
1 Department of Geography, The
Pennsylvania State University
2 College of
Information Sciences and Technology, The Pennsylvania State University
Tool(s):
We use the SensePlace2 tool that was developed at Penn State
as part of the VACCINE project. The tool
has a backend wherein text documents can be stored and indexed. We use open-source software ANNIE and OpenCalais to perform entity extraction on the documents. We adapted these tools to add rules and
dictionaries of place names in the Vastopolis
area. We wrote code in Java to geo-code
the extracted location names. We use
Excel to store the metadata associated with each text document. Lucene is used to index
the text and enable keyword search. We use NodeXL, a
tool built by the Schneiderman group and available
open-source, which can display and analyze a network graph, to examine the
entity-relationship network extracted from the text documents. The networks that were visualized contained
three types of entities: individuals, companies, and other organizations. We used the commercially available ArcGIS
(from ESRI) to derive spatial plots and Excel to draw timeline plots of
selected news articles. We wrote substantial amounts of custom software to
augment these tools. For example, to
augment ArcGIS so that we can build geographical grids, location-related
information, individual-related information, networks among individuals,
etc. We compared occurrence of events in
the spatial plots both using raw frequency and normalized frequency using the
population of the areas. We wrote our
own code to match fuzzy strings to find individual names that match across
datasets. The custom-code written in
Java was built in six man weeks coding by experienced developers.
Video:
ANSWERS:
MC 3.1 Potential Threats: Identify any imminent terrorist
threats in the Vastopolis metropolitan area. Provide
detailed information on the threat or threats (e.g. who, what, where, when, and
how) so that officials can conduct counterintelligence activities. Also,
provide a list of the evidential documents supporting your answer
Threats
Summary
The military weapons lost
included surface-to-air missiles. These
missiles could have been used to bring down an airplane but there may be
additional ones. Imminent use of these
is a major concern. The mass animal
deaths on 4/1, the fish dying around 5/19 point to potential additional
food-borne attacks in the near future, especially, given one was
successful. The theft of computer equipment
points to future cyber attacks. The
suspicious turkey implies that someone tried to test the security system at the
airport and are trying to put a bomb in the cargo to
take an airplane down. The explosion at
the Smogtown chemical plant could be because someone
was trying to make a bomb there. The
bomb could have been planted in a turkey or some plane cargo or could be a
dirty bomb that could spill into a water-way or farm. The stolen university equipment could be used
to make synthetic toxin or other hazardous bio-materials. The immigration officer, people working at Smogtown chemical plant, the computer lab, the bio lab at
the university, and the trucking company could be potential targets to follow
up. Teddy Rao
and Cody Woods, two members of Psycho-Brotherhood escaped from the Center for
the Criminally Insane. Two men who had
escaped from the Center were later caught. These men caught with enough
materials to make several explosive devices.
Gun sales in the local area have gone up. The organization Psycho-Brotherhood should be
investigated.
Weapons Acquisition and Use
2287 4/26/2011 Military Weapons stolen, 20 rifles, AA missiles
4293 4/30/2011 Park shootout
499 5/20/2011 Military weapons discovered at airport.
The major concern is the missing anti-aircraft missiles that could potentially be used to bring down airplanes. The discovery of military weapons at the airport may mean that the weapons are going to be used at a different location. The rifles discovered at the airport should be investigated further as to whether these were the same rifles that were lost or whether these are different ones. Rifles should be easy to check using their IDs.
Prediction: There may be an attempt to shoot down airplanes in the near future, or use as an IED in a crowded area, as modern AA missiles primarily use shrapnel to inflict damage. It could cause mayhem if they are detonated in a crowded area, like a stadium, possibly at a planned public event sometime in the future in the Vastopolis area, where the location of the event, and the route of people traffic is already known, like a parade or a car race (mentioned in a tweet on 5/20).
Illegal Entry
2335 3/31/2011 U.S Immigration officer held in passport ring.
Who did the immigration officer process? Who got the fake passports? This may lead to terrorists who entered the country illegally. There are some tweets that talk about getting passports for immigrant drivers repeated around 5/8 to 5/16. Some investigation of the users is useful although this seems like a minor issue.
Security Test
1243 5/14/2011 Suspicious turkey at airport.
This was likely a test to beat the security system, trying to see if he could get through the system. The turkey was stuffed with some electric device.
Prediction: Airport security should be beefed up. The same article says that there is vulnerability in the system with respect to such items passing checked baggage.
Food Supply Attack and Poisonings.
2385 4/1/2011 Mass animal deaths
2664 5/17/2011 Explosions at Smogtown chemical plant.
1030 5/17/2011 Truck collision
1038 5/19/2011 Fish deaths
Possible attack on the food supply, biological or chemical or both. The first article mentions that there will be immediate effects on the market as the food supply is lower, and prices for consumers will rise. This is followed by the sickness at the river area, pointing to the fact that the river is contaminated in some way, probably the same things that is killing the fish is causing people to get sick.
Prediction: Similar occurences in nearby areas and/or increase/continuation of such events.
Money and Materials Acquisition
926 4/10/2011 Library computer theft
2243 4/13/2011 Conspiracy
1785 4/26/2011 University equipment stolen
2664 5/17/2011 Explosions at Smogtown chemical plant.
1750 5/12/2011 Major robbery attempt fails
This is the set of possible equipment a group may have at their disposal. Both separately and together this represents a potential threat if in the hands of people who know how to use such resources.
Long Term Threats
Conspiracy to
Commit a Chemical or Biological Attack
2335 3/31/2011 U.S Immigration officer held in passport ring.
926 4/10/2011 Library computer theft
2243 4/13/2011 Conspiracy
1785 4/26/2011 University equipment stolen
2287 4/26/2011 Military Weapons stolen, 20 rifles, AA missiles
2664 5/17/2011 Explosions at Smogtown chemical plant.
These events can provide an organized group with access to the country, computing hardware, untraceable funds, molecular biology hardware for manufacture, weapons and explosives for implementation, and chemical materials to be used in the manufacturing process.
Prediction: A second poisoning of the river, or release into the air at some time in the future with far more drastic consequences than the previous poisonings.
We used the following iterative process:
1. Search for interesting keywords like stolen, gun, missile, weapon, death, threat, etc.
2. Extract entities. Plot the entities and their relationships. Plot the interesting events and the titles of the events to obtain spatial and time-line views.
3. Analyze the results and update the keywords and entities to modify the views to examine sub-events. For example, we examined all truck accidents, all fish deaths, animal deaths, equipment stolen separately. Clusters in space-time indicated events.
Figure 1: A view of interesting
events and their locations plotted using ArcGIS. The different colors represent different
two-week intervals: 3-March, 4-4.15-April 1-15, etc. Obtaining a view where both spatial and
temporal correlations and distances can be easily seen provided us with an improved
comprehension of the set of events.
Figure 2: Timeline of events from news articles and microblogs. Black line and grey bars show the number of flu and
diarrhea, vomiting and stomachache cases, respectively.