Entry Name: "TUE-Cappers-MC1"

VAST Challenge 2017
Mini-Challenge 1

Team Members:

Bram C.M. Cappers, Eindhoven University of Technology, b.c.m.cappers@tue.nl PRIMARY

Student Team: YES

Tools Used:

Provide a list of tools used. Examples:

· EventPad developed by Bram C.M. Cappers and Jarke J. van Wijk at the Eindhoven University of Technology, The Netherlands. This system is going to be presented at the InfoVis track of VIS 2017.

· Adobe Photoshop (to compute distances between routes)

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

Approx. 20 hours, 5 hours of analysis, 5 hours on video recording+editing and 10 hours to prepare submission.

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2017 is complete? YES

Video

https://www.youtube.com/watch?v=IBgJ3R9cAvQ

Summary System

In the Eventpad system every glyph sequence (Figure1A) corresponds to sensor events grouped by an attribute of interest. Using regular expressions and predicate logic (Figure1B), we can search and color these glyphs based on event attributes of interest (Figure1C). For question 1, sequences are grouped per car-id per day. For the other questions we either group the events by car-id or by date. We start analysis by creating 5 rules coloring all camping events orange, entrances green, general-gates blues, rangerstops yellow, and rangerbase events pink. All time and date formats are stated in the format “day-month-year hours:minutes” (24 hours notation).

Figure1 A) Event sequences without rules B) Construction of rules using regular expressions and logic. C) Result after rule application.

Questions

1 – “Patterns of Life” analyses depend on recognizing repeating patterns of activities by individuals or groups. Describe up to six daily patterns of life by vehicles traveling through and within the park. Characterize the patterns by describing the kinds of vehicles participating, their spatial activities (where do they go?), their temporal activities (when does the pattern happen?), and provide a hypothesis of what the pattern represents (for example, if I drove to a coffee house every morning, but did not stay for long, you might hypothesize I’m getting coffee “to-go”). Please limit your answer to six images and 500 words.

1. General Patterns

To study frequent daily patterns, we stack visually similar sequences and sort them by frequency (Figure2A). Figure2D shows an alignment of frequent sequences using Multiple Sequence Alignment. This reveals 4 daily patterns (Table1), namely:

Pattern	Frequency	Target Group	Regex Pattern
1. From a camping to an exit	~26%	~50% 2axle cars ~30% 2axle trucks ~20% 3axle trucks
2. From an entrance to a camping	~25%	~50% 2axle cars ~30% 2axle trucks ~20% 3axle trucks
3. Day trips through the reservoir	~43%	All vehicles but 2P ~40% 2axle cars ~23% 2axle,3axle trucks ~15% 4axle trucks, 2axle, 3axle busses
4. Ranger roundtrips	~4%	2P traffic only

Table 1 Frequent patterns in 25523 sequences. Patterns are matched over entire sequences.

Figure2 A) Prototype showing event sequences using glyphs. B) Scented widgets of (derived) attributes. Selections and rules are shown in C) and D).

Surprisingly, the pattern from camping to camping only occurs 14 times. Sequences that do not fit patterns in Table 1 are vehicles driving overnight.

2. Camping patterns

Only 2axle, 3axle cars and trucks, and 2P vehicles drive to camping’s (Figure3A-1) between 05:00-01:00 (Figure3A-2). Extraction of camping events (Figure3B) shows that camping1 is least visited (Figure3C-3) and not visited in December, March and April (Figure3C-4).

Figure3 A) Searching for camping events. B) Inspecting trucks sequences. C) Grouping sequences by length.

3. Day trips

Figure4A shows bus and 4axle truck traffic. These vehicles do not visit camping’s.

Figure4 A) 4axle trucks and bus patterns. Bus patterns are selected. B) Application of detailed ruleset. C) Alignment of bus and truck traffic.

Bus sequences (selected sequences, Figure4A) only visit the reservoir once per day. The last sequence represents unauthorized access of a 4axle truck (see question 3). Figure4B also shows this pattern along with sequences that are still driving in the area (Figure4C-1)

2.1 Traffic speed

Figure5 A) Sequences from and to an entrance. B) Travel time between routes. C) Count route lengths using Photoshop.

28% of the sequences visit camping’s (Figure5A). These vehicles are 2axle, 3axle cars and trucks (Figure5A-1) and are inactive between:

· 02:00- 04:00 (Camping’s are probably closed) and

· 20:00-21:00. (Probably dinner time)

The rest of the traffic travels directly from entrance to exit. Figure5B shows that direct routes from E0 to E3 on average takes 1341 seconds. Using Photoshop histograms we know that the travel distance is 219 pixels (Figure5C). This corresponds to:

.

The speed between E2 and E4 does not exceed the 25 mph limit.

4. Ranger shifts

Figure6 A) Frequent ranger shifts. B) Extraction of rangerstops. C) Properties of longest shift.

Rangers do not travel between 04:00-05:00 and always start/end in the rangerbase (Figure6A). Figure6A-1 shows that 62% of the ranger shifts visit camping’s between 06:00-22:00. Rangers pass camping8 in half of these shifts (Figure6B-2). Rangerstop6 is most popular, since it is the fastest way to get from the west side back to the base (Figure6C-3). The longest shift is approximately 9 hours visiting almost all stops (Figure6C-4).

2 – Patterns of Life analyses may also depend on understanding what patterns appear over longer periods of time (in this case, over multiple days). Describe up to six patterns of life that occur over multiple days (including across the entire data set) by vehicles traveling through and within the park. Characterize the patterns by describing the kinds of vehicles participating, their spatial activities (where do they go?), their temporal activities (when does the pattern happen?), and provide a hypothesis of what the pattern represents (for example, many vehicles showing up at the same location each Saturday at the same time may suggest some activity occurring there each Saturday). Please limit your answer to six images and 500 words.

We group the data by car-id to find activity over multiple days. Table2 shows the travel time distribution of vehicles in the data set.

Car-type	Duration distribution (logarithmic)	Normal	Outlier’s
2axle cars		Between 0.5 hour and 37 days	≈2 minutes and 350 days
2axle trucks		Between 0.5hour and 20 days	≈4minutes and 107 days
3axle trucks		Between 0.5 hour and 10 days	≈4minutes and 23 days
2P vehicles, 4axle trucks, and busses		Between 0.5-10 hours	≈12 hours

Table 2: Travel duration of vehicles in the reservoir.

Extracting enter and leave events in the data shows that almost all vehicles enter and leave the reservoir at most two times. The longer sequences can directly be related to different vehicle types (Figure7C)

Figure7 A) Event sequences are grouped by reservoir visits (B). C) Result after applying rules.

Figure8 Weekly pattern of tourist bus in reservoir. B) Geographical representation of tourist bus route. Route in red is the place where the bus entered the reservoir.

In case of the two axle trucks there is a sequence of 107 days showing repeated activity. This truck made trips from camping 4 to entrance between 23:00-00:00 and 14:00-15:00 (Figure8A-1). This is probably a tourist bus, since it only rides during high-season of 2015 (July –October, Figure8A-2). For unknown reason, the bus only travels on Monday, Sundays, and Fridays (Figure8A-3).

Figure9 A) Weekly pattern vehicle visiting camping. B) Geographical representation of taken route.

There is also a sequence of 37 days (Figure9A-1) in which one 2axle car (Figure9A-2) travels every Sunday and Fridays (Figure9A-3) between 24-06-2015 until 30-07-2015 (Figure9A-4) from Entrance 0 to camping 6 and vice versa. The vehicle only travels between 13:00-14:00 and 22:00-23:00 (Figure9A-5). On average he spends 2.3 days in the camping before he leaves the camping. Maybe this tourist has a stationary camper located there.

Figure10 A) Long stay of visitor at camping 0 B) Graphical representation of the driven pattern.

Another 2axle car traveling on Fridays on Sundays is depicted in Figure10A-1 and Figure10A-2. He travels from entrance3 to camping0 (and vice versa) between 09-03-2016 until 22-04-2016 (Figure 10A-3). He also spends approximately 2 days at the camping before he leaves.

Night patterns

We inspect patterns in night traffic by filtering the traffic between 23:00 and 04:00 (Figure11A). This reveals that only camping4 was entered in this period (Figure11B-1) and most of the 4Axle trucks and busses drive during the night (Figure11B-2).

Grouping the sequences by date (instead of car-id) and coloring the events by car type shows the different type of vehicles that drive per day in the reservoir. Sorting the sequences alphabetically shows that on certain days car type events occur more than others (Figure11C). Figure11C-1 shows that night activity of 3axle busses mostly happens during low season (i.e., the period October-April).

Figure11A) Query for all traffic over night. B) Large trucks and busses drive in the night. Few camping’s are visited. C) Overview of night patterns per day.

3 – Unusual patterns may be patterns of activity that changes from an established pattern, or are just difficult to explain from what you know of a situation. Describe up to six unusual patterns (either single day or multiple days) and highlight why you find them unusual. Please limit your answer to six images and 500 words.

Figure12A colors events by the car type. Sorting the sequences alphabetically shows that every vehicle has frequent short and rare long sequences. Selecting the longest sequence of 3axle trucks shows that it only visited camping 1 and 5 (Figure12A-1) on September 8 and 9 (Figure12A-2) between 08:00-09:00 and 16:00-18:00 (Figure12A-3). The long 2axle truck sequence in Figure12B corresponds to the tourist bus in question 2.

Figure12A) Frequent and rare car patterns. B) Overview widget shows a long pattern in 2axle truck sequences.

1. Long term visit

Figure13 Strange pattern of a vehicle traveling through the entire reservoir

The third most frequent car-id (Figure13-1) corresponds to a 2axle car traveling from 06-06-2015 to 20-05-2016 on days other than Wednesday and Saturday (Figure13-2). He travels from entrance0 to all the camping’s (except for camping7 and 8) between 08:00-19:00 (Figure13-6). After approximately one month (Figure13-3) he travels to the next camping (except in the month April, Figure13-4). According to the last event, he never left the reservoir (Figure13-5). The ordering of camping visits is strange and seems random. Maybe he is an ornithologist seeking for the bird species.

2. Unauthorized passage

Figure14 A) Search for rangerstops B) Deviating pattern in overview C) Inspection shows the presence of a 4axle trucks in range routes. D) Graphical representation of taken route.

There are 23 cases where 4axle trucks (Figure14B-2) are going from rangerstop6 to 3 to 6 via gate6 (Figure14A-1). The sequences occur between May 2015 and May 2016 (except April, Figure14B-3) between 02:00-05:00 (Figure14B-4) on Tuesdays and Thursdays (Figure14C-5). Maybe some construction material has to be delivered there.

3. Traveling truck

Figure15 A) Long stay visitor in 3 axle truck. B) Route in red is only traveled once.

Between 12-07-2015 and 04-08-2015 there is a 3axle truck (Figure15A-1) traveling between camping6 and entrance4 (Figure15A). He initially entered the reservoir via entrance2 (Figure15A) and drives only on Tuesdays and Sundays (Figure15A-2). It is unclear why he entered the reservoir from Entrance2.

4. Very short visits/Day-trip Loops

When extracting daytrip patterns in question 1, we noticed some sequences whose entrance and exit are the same. Table 3 shows a summary of these “loops”.

From	To	Duration (seconds)	Arrival Date
E1	E1	5	04-07-2015 22:02
E1	E1	5	23-03-2016 21:06
E2	E2	5	26-06-2015 22:34
E3	E3	6	01-09-2015 20:45
E3	E3	5	18-05-2016 18:10
E0	E0	5	22-102-2015 20:03

Table 3 Routes with the same entrance and exit.

Inspecting these sequences shows that they are caused by 4 axle trucks (Figure17D-1), traveling on all days (except Monday and Friday, Figure17D-2) between 18:00-22:00 (Figure17D-3).

Figure17 A) Extracting direct roads B) Inspecting routes with a more detailed ruleset C) Grouping traffic by car type. D) Histograms showing the location of these sequences in time.

The sequences could be delivery trucks dropping supplies at reservoir entrances for special occasions (e.g., fireworks on the 4^th of July). Figure17D-4 however shows that the time between entering and leaving is at most 5 seconds.

4 –– What are the top 3 patterns you discovered that you suspect could be most impactful to bird life in the nature preserve? (Short text answer)

1. Vehicles on the road between entrance 0 and 3 drive too fast. The noise can disturb the wildlife.

2. The repeated access of vehicles to unauthorized locations (in the middle of the night) and the presence of systematic travel activity across the entire reservoir (e.g., tourist busses) during high-season can prevent wildlife from establishing a proper breeding place.

3. The continuous nightly activity of vehicles such busses and trucks over the entire year can disturb the wildlife.