Entry Name: "USTUTT-Thom-MC3"
VAST Challenge 2014
Mini-Challenge 3
Team
Members:
Dennis Thom,
University of Stuttgart, dennis.thom@vis.uni-stuttgart.de PRIMARY
Michael Wörner, University of Stuttgart, michael.woerner@vis.uni-stuttgart.de
Steffen Koch,
University of Stuttgart, steffen.koch@vis.uni-stuttgart.de
Student Team: NO
Team
Number: 39
Streaming
User ID: dennis.thom@vis.uni-stuttgart.de
Analytic
Tools Used:
Components of Scatterblogs,
developed by the University of Stuttgart and adapted
for the challenge. Weka for the hierarchical clustering.
Approximately how many hours were spent
working on this submission in total?
30
May we post your submission in the
Visual Analytics Benchmark Repository after VAST Challenge 2014 is complete? NO
Video:
Video is provided here
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Questions
Please note - this challenge
contains a question that is time-dependent.
Within 3 hours of starting the final data stream, send an email to VASTChal2014MC3@vacommunity.org containing your answer to question MC3.1. Please include a copy of your answer to MC3.1
in your final answer form also. Your answers to MC3.2 and MC3.3, along with
your video, are due July 8.
The
responses to these questions should be incrementally built, as you (the
contestant) acquire information from each streaming data segment you
receive. Your submission will answer these questions in consideration of
all of the streaming data segments.
MC3.1 -
Within 3 hours after start the final data stream, send an email to VASTChal2014MC3@vacommunity.org containing:
a.
An image showing the streaming data in
your visual analytics tool. In this image, identify an event of interest that
you intend to investigate further.
b. The content of the final message in the data stream
The
image shows our tool doing the real-time analysis. Here our message-classifiers
have identified several messages highlighting an explosion connected to the
dancing dolphin fire that was reported earlier in the messages. As the fire was
reported to be under control, we suspect that the explosion might be an act of
terrorism and could be connected to the missing GAStech
employees. Therefore we will investigate this further and try to find more
suspicious activities connected to the dancing dolphin fire.
Last
message in the stream was: 21:32:00 gardener4958 RT @KronosStar
There has been an explosion from inside the apartment building. Several people are down. #KronosStar
#DancingDolphinFire #AFDHeroes
MC 3.2 - Describe the timeline of up to
five major events that you discover in the streaming data. This timeline should
include information from all three segments of the data stream if needed.
Use specific microblog records and call center data to support
your description, but do not simply mimic back the data stream. Provide a
concise description of important participants, locations and durations.
Focus your response on the events themselves, rather than on the individuals
reporting the events. Please limit your answer to no more than ten images and
1500 words.
To
solve this challenge we developed a novel streaming-enabled visual-analytics
system that is supposed to be employed in a two-display setup.
The
first display provides an overview
of streamed data and is composed of following components: A) A table showing
message details such as user-id, time, place and text B) A temporal overview of
message volumes. The system performs a real-time sentiment analysis and the
colors show the respective volumes of positive (=green), negative (=red) and
neutral (=blue) messages. The component can furthermore be used for 2-step
temporal filtering (selecting a timeframe in the upper part shows the data in
more detail in the lower part). C) An LDA based overview that shows topics of
messages in the current user-selection. D) A map overview of gelocated data. Red dots show regular messages. Blue dots
show ccdata. If geolocation is not explicitly given,
we try to find location-names in the text and estimate possible locations. E) A
filter-flow graph that can be used to store and combine filter configurations
using boolean operations. F)
A control bar including textual search and means to activate geo-based tag-map
visualizations.
The
second display provides a detail view
of data based on an explorable topic structure. Once
the user has selected a subset of messages in the overview-display based on
temporal filters, textual filters or combinations of both, he can send them to
the detail display, by clicking the “Cluster Tweets”-Button in the control
panel (G). Based on a hierarchical clustering the underlying topic structure of
the subset is then extracted and visualized in a treemap-display
of representative tagclouds (H). The user can explore
this hierarchical overview using the mousewheel,
which changes the display layer of the tree hierarchy, and by selecting nodes,
which shows the individual tweets contained in a sub-tree of the hierarchy in a
table. Selected nodes can be further zoomed-in by clicking the zoom-buttons in
the control panel. In this case, the minimap (I)
shows the current viewport within the complete tree structure.
Using
the tool we were able to detect the following events: A POK rally, a fire in the
Dancing Dolphin apartment building, two
traffic accidents with subsequent pursuit through the city, a shooting, and an explosion in the burning apartment building.
POK rally: Shortly after
17:00, a POK rally with an estimated attendance of 1200 people starts in Abila
Park. While streaming the data one could immediately identify this event as a burst
of data is evident. Further investigating the connected messages using the
hierarchy in the detail view illustrates how the event proceeds and which
speakers and activists participate. Some surrounding streets are closed by the
police. @AbilaPost covers the event in several
messages and names speakers Sylvia Marek, Dr. Audrey McConnel
Newman, Lucio Jakab, and Prof. Lorenzo Di Stefano.
The rally remains peaceful but with strong police presence and ends after a
one-hour concert of Victor-E at about 20:15.
Dancing Dolphin fire and explosion: At about 18:40, an intense fire breaks out in the
Dancing Dolphin apartment complex in the east of the city. @dangermice
is an eye-witness that regularly posts news over the next three hours. People
are injured and nearby buildings are evacuated. The fire is reported to be
under control by @NewsOnlineToday at 20:30, but shortly
after, the building partly collapses. Additional fire trucks are dispatched and
the fire department focuses on saving neighboring buildings. At 21:30, an
explosion occurs, injuring more people.
Traffic accidents and shooting: At about 19:20, a black van
hits the car of @hennyhenhendrix on Souliou street not too far from
the Dancing Dolphin complex. The van flees the scene, hits a bicyclist on Ipsilantou Avenue and continues west. Police pursue the
vehicle and stop it near the intersection of Alexandrias
and Ithakis. Shots are exchanged, trapping people in
the surrounding businesses. Some of them (@megaMan, @tuccotucco, @Simon_Hamaeth) post
their observations from inside these buildings. A police officer is injured and
taken to the hospital. A SWAT team is dispatched and arrives at the scene at
about 19:55. After lengthy negotiations and an apparent argument inside the
van, two suspects surrender at about 21:15. Two women, presumed hostages, are
rescued from the van.
MC 3.3 – Select one of your five major
events from question MC 3.2 that you consider to be most likely to provide
additional clues to the investigation of the GASTech
disappearances. Describe the roles of the participants. Describe how
other events you identified in MC3.2 may have influenced your selected event.
Provide a hypothesis and evidence as to whom you suspect as being directly
involved in the GAStech disappearances, either as
perpetrators or victims. Please limit your response to no more than five
images and 500 words.
We
suspect that the hostage situation and the shooting with the drivers of the
black van are most likely connected to the GASTech disappearences. Clearly the drivers of the van transporting
the hostages might have been distracted or involved in a fight within the car.
Thy hit another car and the bicyclist and flee in panic. To follow the van we
create a filter module in our filter-flow graph.
Based
on the recognized messages and the ccdata information we can follow the path of
the van until the police pursuit ends at the Gelatogalore
corner. In the subsequent course of events the shooting begins, the drivers try
to work a deal with the police and finally surrender. Several witnesses and
commenters suspect that the hostages – two females - are two of the missing GASTech employees. As there are no other reported
kidnappings during the time, we follow that hypothesis.