Entry Name: NYU-Bowen-GC

VAST 2015 Challenge
Grand Challenge

Team Members:

Bowen Yu, New York University, yubowenok@gmail.com PRIMARY

Bo Zhou, New York University, bz387@nyu.edu

Student Team:

Yes

Did you use data from both mini-challenges?

Yes

Analytic Tools Used:

Ad-hoc System (self-implemented for VAST challenge 2015)

Approximately how many hours were spent working on this submission in total?

233

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2015 is complete?

Yes

Video Download

Video:

https://www.dropbox.com/s/5qftbw4waw24u46/NYU-Bowen-GC.wmv

-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Questions

NOTE:

- We have re-assigned IDs to the individuals based on their order of appearances throughout the weekend. The individuals are then grouped based on the similarity between their movement trajectories. Group IDs were assigned.

- Our system displays the newly assigned IDs. Therefore the snapshots all contain our custom group IDs.

- However, the person-ids listed in the answer use the IDs from the raw data.

For each of the following questions, consider both the movement and communications data.

GC.1 – Scott is not a paying customer and does not have an ID. Describe Scott Jones’ activities in the park during the three-day weekend. Who does he spend most of his time with? When does he arrive? When does he leave? What route does he follow?

Limit your response to no more than 10 images and 1000 words.

It is easy to identify that, Scott Jones has two shows every day, at facility 63 Grinosaurus Stage. We have found there are 8 persons, who are probably Scott’s staff for the shows. Their IDs are: 1080969, 644885, 1781070, 521750, 1787551, 1935406, 1629516 and 1600469. Their move path are shown in the image below. In the mornings they come at 8:45am, use the same path move to facility 63 the Stage half an hour before the shows. The shows are from 10am to 11 am. They start moving at 11:30am and follow the reversed path that they come to leave the park. In the afternoons, they come again at 1:45pm. They take the same path, arrive at the Stage at 2:30pm. The afternoon shows are from 3pm to 4pm. The start moving at 4:30pm, they left the park at 5:15pm.

The show was cancelled on Sunday afternoon. So they didn’t come then.

At the beginning of Friday around 9am, Facility 4, Galactosaurus Rage, had far more visitors than any other facility. The number was significantly higher than its riders in the next two days. The first image is the chart of number of people at Facility 4 on Friday, the second one is on Saturday. We think there could be an opening ceremony event there. It is very possible Scott was there during that time, so everybody went there in the morning.

Besides, it can be seen that Facility 3, Auvilotops Express (Thrill Rides), had far more visitors on Sunday (over 600 at peak). The two charts below show the people flow at Facility 3 on Sunday and Friday respectively. We think that probably Scott were there in the afternoon for a long event (such as a closing ceremony). Alternatively, it could be due to some emergency actions taken after the crime, so that people were guided there.

We have also tried to find other patterns that can tell if Scott have other activities during the three days. For example we have tried to track everyone that has watched a show. Almost everyone leaves the show exactly on it finishes (11am and 4pm). It turns out they move quite randomly, so we excluded the possibility that he had activities somewhere else after the show. The image below is the heat map of numbers of people that left from the show.

The image on the right illustrates the positions of people that have watched the show on Sunday morning, at 8 minutes after the show finished.

We have also tried to look at the number people at each single facility, as well as the messages sent from each facility throughout a day, in order to find if there is something important going on. Unfortunately, we haven’t found any pattern that we think are guaranteed to relate to Scotts’ activity.

There was a group of four people, ID: 1690685, 629048, 1797150 and 1486047. They watched the both the two shows on Friday. See image below. They stayed checked-in at the Stage during the 4 hours between the two shows. Though weird, there is no proof that they are related to Scott.

There was another group of three people, ID: 1856633, 631146 and 802410 on Saturday. They stayed one more hour in the Stage facility after the show. Then they left the park. One of them also came on Sunday morning and watched the show again. We guess they are Scott’s fans. They stayed late maybe because they were talking with him.

GC.2 – Identify up to 8 issues with park operations during the three-day weekend. Provide a rationale for your answers.

Limit your response to no more than 8 images and 800 words.

1) Queueing at entrance.

We have noticed that the queueing time is significantly longer at the main (North) entrance, especially on Saturday and Sunday mornings, visitors wait 20 to 30 minutes to get in the park. However, at the same time, the other two entrance are very light loaded, people just waited 1 minute to get in. The park should balance the load, let the visitors use the other two entrances to enter.

2) Unbalanced restroom usage.

We found the restroom usage is very unbalanced among different restrooms in park. The image below shows usage the three restrooms: 51, 49 and 66 on Saturday. The first row is restroom 51, which locate in the center of the park. It is always heavy loaded all the time. The peak usage could be 110 people using it at the same time! The second row is the other one which is located besides the left entrance of the park. It has about less than half of the usage as 51. The third row is the one besides the show Stage. We can observe two peak at 11am and 4pm. It is due to the Scott’s show finished at that times. People need to queue for a long time waiting to use this restroom.

3) Visitor staying overnight.

We found a person (ID: 657863) who stayed overnight on Friday night, who left the park in the morning of Saturday. This person was with his 4 friends on Friday night. They might be drunk because they had gone to the beer garden. His friends left the park but left him at facility 20. We don't know if it was a prank or accident. But the park should definitely prevent this from happening.

4)
Place the facilities according to their types, to help visitor utilize their time better. Some people travels all around to play the rides, especially those who loves Thrill rides. Because many thrill rides are distributed at the boundaries of the map, unlike the other facilities. For instance, the kiddie rides are put together.

5) Improve app quality.

We have detect several abnormal movement in the data. We think it was because of some bug in the app or signal problem. The strange movement including: suddenly moving from one side of the map to the other; jumping positions of a single ID; disappearing in the restroom and reappear after 5 hours. They are more detailed described in our MC1 answer sheet.

The facility 62 Liggement Fix-me-up has been visited frequently during the day. Especially on Friday, there were around 40 people at 2pm need help there. We suggest the park to improve their app quality and make visitor happily enjoying the park, not fixing the troubles.

6) Add medical center facilities.

Some visitors didn’t enjoy the park very well. They left so early. Some of them played several Thrill Rides. Our assumption is they felt ill after doing too many of them. Maybe they need medical help.

7) Improve the security:

a. For those facilities that need visitors to check-in, enforce them to do so.

In the morning of Sunday at facility 32 the Pavilion, when and where the crime happened. A large group of people get in and didn't check in. There were also some people walking around the park all day, but didn’t check in once except at the entrance. The first char in the image below is facility sequence chart, which depicts the visiting sequence of selected visitors. The second char is the facility people flow which shows the number of visiting people at a selected facility. If visitors have checked in, they are painted in darker color. We can see the security at the Pavilion is really bad. Many visitors didn't check in and got in there. These people are really closed to the crime there, which will be explained later in answer to Q3.

b. Analysis and watch over and suspicious pattern of visitors.

Again, look at the first chart in the image above. There are several visitors with really strange visiting pattern in common, which are highly suspicious:

· Everyone is by himself/herself.

· Each of them only shows up for a single day.

· They spend lots of time stay around several facilities of ‘Rides for Everyone’ or ‘Shows & Entertainment’. But they NEVER check in.

· They don’t eat food, they don't EVEN go the restrooms. Even for the entire day staying at the park. The just roaming between some particular kinds of facilities.

· Some of them go the facility 32 Pavilion more than once. It seems they are fans of Scott Jones. But they actually never watch Scott’s show on the stage.

We highly suspect they are responsible for the crime. The security of the park should pay more attention in strange patterns like this.

8) Ensure smooth run of rides.

We found Facility 20 (Scholtz Express, Rides for Everyone) might have stopped running on Saturday. The park should make sure rides are more RELIABLE.

GC.3 – For the crime, describe the following, and provide your rationale:

a. When did the crime occur?

b. Where did the crime take place?

c. Who are the most likely suspects in the crime?

Limit your response to no more than 5 images and 500 words.

a. The crime occurred between 10:30am to 10:40am on Sunday morning.

b. They crime took place at facility 32 Creighton Pavilion.

c. ID: 921888 is most like suspects that fulfilled the crime. There are some other suspects that may helped him, their IDs are: 159893, 430595, 1711922, 1601276, 1658667, 500084, 1149884, 47441, 655378, 1563594 and 1269018.

These people have really strange visiting patterns, described in previous answer.

The picture below shows these patterns. The red targets are this type of visitors on Saturday. The left top table is the Facility Percentage shows the percentage of time one spend in each type of facilities. The bottom chart is the visiting sequence of facilities. The colors indicate the types. The darker colors tells if one has checked in at the facility. The visiting time for Pavilion (in purple) is about 10 to 20 minutes.

Besides the 5 suspicious patterns mentioned in MC1.1. Some of them also frequently visited facility 32 the Pavilion where Scott’s items were exhibited, in the morning of Sunday. Some of them left the park very soon after the vandalism was discovered; some of them even frequently visited and stayed near the scene in the afternoon, after the park operations closed it.

We guess the reason they were roaming around on Friday and Saturday is they were looking for the moment of least visitors at the Pavilion in three days. Although they were on different IDs, but we still think they are suspicious, they may have used fake IDs to get different logins.

In our hypothesis, the crime was first discovered at 11:30am on Sunday by some members of a large group. The group contained 37 people. They move with the same pattern in MC1 data. The group members’ person-ids include: 1038892, 1041478, etc. After some of them discovered the vandalism, they first communicate using the app within their group.

Then they report it to the Information Center (ID: 839736), maybe some of them started calling the police, or posting Tweets etc. (send via ID: external). Some other groups of visitors (such as a 10-people group, including ID: 1559001 etc.) joined them and started send out messages. The corresponding pattern can be found during 11:45am to 12:00pm, on the volume chart of ID: external. We can also learn from the below image, this 10-people group is a big fan of Scott. They just watched his show in the morning. They were angry at the crime and immediately started telling everyone what happened.

Around 12:00pm, everyone in the Pavilion knew what happened. We guess it was because the park operations announced that: “Dear, visitors! We are sorry for the inconvenience! The Greighton Pavilion is going to be closed in 30 minutes!”

“Why?” “What happened?” many visitors who were still in the pavilion immediately asked the Information Center. This was the direct cause of sudden peak of receiving messages of Information Center (ID: 839736).