Entry Name: "UMD-Yalcin-MC1"

VAST Challenge 2015
Mini-Challenge 1

Team Members:

Adil Yalcin, University of Maryland, College Park, yalcin@umd.edu PRIMARY

Matthias Nielsen, Aarhus University, matthiasnielsen@cs.au.dk

Student Team: YES

Did you use data from both mini-challenges? NO

Analytic Tools Used:

· Tableau

· Keshif (http://www.keshif.me) - developed at HCIL, UMD.

· Keshif browsers used for this challenge is available at: http://cs.umd.edu/~yalcin/DinoWorld

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

About 15 hours total

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

Video Download

Video:

http://youtu.be/n-35Xc9s7CI

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

Questions

MC1.1 – Characterize the attendance at DinoFun World on this weekend. Describe up to twelve different types of groups at the park on this weekend.

a. How big is this type of group?

b. Where does this type of group like to go in the park?

c. How common is this type of group?

d. What are your other observations about this type of group?

e. What can you infer about this type of group?

f. If you were to make one improvement to the park to better meet this group’s needs, what would it be?

Limit your response to no more than 12 images and 1000 words.

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

Of 11374 total visitors {1}, we detected visitor groups by the number of rides, the first and the last check-in per visitor.

We detected 317 groups that include 7 visitors or more (right panel) {2}.

In this browser, length of all the bars visualize the number of unique visitors on that ride.

Based on the same figure, there are 31 groups with 28 or more visitors {3}.

Then, the group size suddenly drops to 11 or less visitors.

The large groups are potentially student groups. Or maybe the group pricing mentioned in the website starts with 28 people?

We noticed that the large groups arrive after 9:00, stay longer and mostly leave after 21:00. {4}

The figure above also highlights visitors on Sunday (black bars) and Saturday (orange bars).

While there are 3.9k visitors who came in both Saturday and Sunday {5}, the large groups only visited on a single day.

That is, there is no GroupID with both black and orange lines.

In contrast, there are groups of size 8 or less that came on both days, shown in our submission video.

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

Grouping (selecting) visitors by the number of check-ins:

Next, we focus on visitors who had high (and low) number of check-ins, and analyzed the distinct rides they took.

The screenshot on the left shows the popular rides/locations for visitors with at most 10 check-ins (819 visitors).

The screenshot on the right shows the popular rides/location for visitors with at least 25 check-ins (1260 visitors).

While the popularity order of the ride changes between visitors with few and many check-ins,

the differences in the number of visitors for each ride are negligible.

We observed no significant differences in check-in behavior between visitors with few and many check-ins.

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

Suspecting that the crime took place in Pavilion, Sunday 11am (analysis to follow), we tried to find a group of visitors who visited Pavilion and then left the park right after without any ride.

We found one such group of 6 people, shown below.

They checked-in to the pavilion at 11:55, and then left the park at 1:13 without getting on any other ride.

We think this is a susceptible behavior.

We also mapped their exit path from the Pavilion using Tableau:

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

********************************************************************

********************************************************************

MC1.2 – Are there notable differences in the patterns of activity on in the park across the three days? Please describe the notable difference you see.

Limit your response to no more than 3 images and 300 words.

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

40% of check-ins were on Sunday (130k/323k).

The figure below shows the percentage of check-ins on Sunday per ride.

While most rides are at ~40% on Sunday, Creighton Pavilion stands out with its lower share (14%). {1}

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

Now, the check-ins at Creighton Pavilion are focused, and the numbers and bar length show the count of check-in events.

Check-ins on Saturday are orange, check-ins on Sunday are black.

There is no checking in the Pavilion after 11am on Sunday. {1}

This suggests that the theft was noticed at 11:00-12:00 Sunday.

The Pavilion is closed at 10:00 and 15:00 daily for preparation of shows. {2}

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

When you focus on check-in activity per hour and the distribution over three days, Saturday has a steady ~35% of check-ins at each hour.

This is except at 23:00 {1}, where 93% of the check-ins (85 of 91 total) were on Saturday.

The check-ins at 23:00 on Saturday night are listed on the left.

The locations marked WetLand-X are not rides, and these locations are arguably near the theft at Pavilion the next day.

This may relate to the theft on Sunday.

********************************************************************

********************************************************************

MC1.3 – What anomalies or unusual patterns do you see? Describe no more than 10 anomalies, and prioritize those unusual patterns that you think are most likely to be relevant to the crime.

Limit your response to no more than 10 images and 500 words.

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

The following figure shows 328k check-ins.

Black bars show Thrill Rides, orange bars show Rides for Everyone, the chart width shows number of check-ins.

The most popular rides are all Thrill Rides (ex: Galac. Rage), and the following popular rides are Rides for Everyone (ex: Dyke. Thrill).

The strict separation in ride popularity per ride type is unusual.

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

Another anomaly is that there are no significant interaction patterns between the check-in pairs.

In other words, taking ride A does not imply a stronger relation with taking ride B overall.

The figure below shows the number of people taking each ride (on the right), as well as the ride-pairs (the matrix view).

Through a uniform distribution, the pairs of popular rides also have a larger population, as shown with larger circle sizes on the upper right matrix corner {1}.

There is no visitors who used multiple entrances (North, East, West), suggested by the lack of intersection {2}.

We can also expect that the visitors who are riding Kids Rides should be more likely to ride other kid rides.

Sauroma Bumpers is a kid ride (selected in the screenshots below), yet the 3.1k visitors who rode this ride seem to get on all other rides, including Thrill Rides such as Galac. Rage (1.9 visitors).

Maybe the park doesn’t have any policing for kids to not get on other rides?

The right side of the figure shows the percent of the visitors who got on Saurome Bumpers per each ride.

There is a slight correlation between riding kid rides, as 40% ratio among kid-rides is pronounced compared to the ~28% ratio among other rides.

We have not observed differences of distribution among other ride types.

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

In the figure below, the movement patterns are visualized on the park map, with lighter color representing more visitors on that spot through the weekend.

It is unusual that the overall movement patterns are focused around location 50 (restroom, Tar Pit Stop, lower section) and 21 (Carousel, near middle section, 20th ride in terms of on popularity).

The food vendors (yellow blocks near the center, 45-38-41…) have little action, which is unusual since eating is a common activity which would generate more movement patterns.