Entry Name: ASTRI-Tse-MC2

VAST Challenge 2014
Mini-Challenge 2

Team Members:

Carlos K. F. TSE, ASTRI, carlostse@astri.org

Yang LIU, ASTRI, yangliu@astri.org

Mengte MIAO, ASTRI, mtmiao@astri.org

Hua CAI, ASTRI, caihua90@gmail.com

Ji Hyoun PARK, ASTRI, jhpark@astri.org

Amir S. Y. LIAO, ASTRI, amirliao@astri.org

Minjing MAO, ASTRI, minjingmao@astri.org

Xiuhua WANG, ASTRI, waixiu@hotmail.com

Kent K. H. WU, ASTRI, khwu@astri.org

James Zhibin LEI, ASTRI, lei@astri.org

Student Team:

No

Analytic Tools Used:

Python

SQLite3

Excel

ArcMap

CanvasJS

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

80

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2014 is complete? Yes

Video:

http://210.3.53.50/vast/astri-tse-mc2-video.wmv

astir-tse-mc2-video

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

Questions

MC2.1 – Describe common daily routines for GAStech employees. What does a day in the life of a typical GAStech employee look like? Please limit your response to no more than five images and 300 words.

We use CC software to show the routine pattern for GAStech employees. We can select single day or multiple days for one or multiple users to see the routine pattern.

As shown in Figure 2.1.1, in working days, the GAStech employees normally leave their home around 7:30 am. Before going to work, they drive a few minutes to have breakfast. After breakfast, they drive to the company and arrive there around 8:00 am. In the noon (11:00 am to 12:30 pm), they leave the company for lunch, usually at restaurants nearby. After lunch, they go back to work at around 1:30 pm. Later, they are generally released from the company around 5:30pm, and most of them often go to a cafe or a bar to spend a happy night from 6:00 to 22:00 pm.

Figure 2.1.1 Routine pattern of a working day Jan. 17 for all the employees

Some of the employees drive out for dinner/shopping and some stay at home in weekdays, as shown in Figure 2.1.2.

Figure 2.1.2 Routine pattern of a weekend day Jan. 19 for all the employees

In addition, the normal routine pattern can be observed in the view of one employee Isia Vann for all the days (Figure 2.1.3), one normal working day (Figure 2.1.4), and one normal weekend day (Figure 2.1.5). In Figure 2.1.3, it is easy to see that there are some abnormal patterns indicating car driving in the very early morning.

Figure 2.1.3 Routine pattern of all working days and weekend days for one employee Isia Vann

Figure 2.1.4 Routine pattern of one normal working day for one employee Isia Vann

Figure 2.1.5 Routine pattern of one normal weekend day for one employee Isia Vann

MC2.2 – Identify up to twelve unusual events or patterns that you see in the data. If you identify more than twelve patterns during your analysis, focus your answer on the patterns you consider to be most important for further investigation to help find the missing staff members. For each pattern or event you identify, describe

a. What is the pattern or event you observe?

b. Who is involved?

c. What locations are involved?

d. When does the pattern or event take place?

e. Why is this pattern or event significant?

f. What is your level of confidence about this pattern or event? Why?

Please limit your answer to no more than twelve images and 1500 words.

First, we find unusual traffic pattern and the related employees. Secondly, we select the related unusual events in consumptions related to the most suspicious employees. Finally, we analyzes the abnormal patterns for events related to the kidnapping.

For the consumption records, we combine the data from loyalty card and credit card.

2.2.1, Loreto Bodrogi has driving records early in the morning (Jan 7, 2014 03:20:01 am to 03:35:10 am, and from Jan 9, 2014 03:20:01 am to 03:32:43 am) by reading the pattern map. The start points and end points for the two unusual driving records in the location map. Comparing the locations to the tourist map, Jan. 7, he visited Spetson Park. On Jan. 9, he visited Taxiarchon Park. It is important because people will not go out at early in the morning unless they feel it is very necessary. We are confident that it is related to the missing people since it happens twice.

2.2.2, Isia Vann has multiple driving records early in the morning (Jan 07 1:10:01, 3:25:00 and Jan 11 from 3:25:01 to 3:35:00) by reading the pattern map. The end point of driving records in the location map, Jan. 7 is Spetson Park. With the same reasons listed in 1, it is important and we are confident it is related to the people missing.

2.2.3. Hennie Osvaldo has multiple driving records early in the morning (Jan 11, 2014 03:23:01 and Jan 14, 2014 03:20:01).

He drove from the green point to the red point in the mid-night twice

With the same reasons listed in 1, it is important and we are confident it is related to the people missing.

2.2.4, Minke Mies has multiple driving records early in the morning (Jan 09, 2014 03:30:01 and Jan 14, 2014 03:30:01)

Minke Mies drove from the left green point to the right green point on Jan 9 and from the right green point to the left green point on Jan 14.

With the same reasons listed in 1, it is important and we are confident it is related to the people missing.

2.2.5, Possible very early morning meeting of Loreto Bodrogi and Isia Vann at Jan 07 around 3:35 am, in Spetson Park. The calculation of the distance of the two people’s nearest location data is zero. Since Jan. 07 is a working day, it is important to investigate why the two colleagues have to meet at the early morning instead of seeing each other at company working time. Due to the data absence of Isa Vann at 3:35:10, the meeting time and location is a best guess we have.

2.2.6, Possible very early meeting of Loreto Bodrogi and Minke Mies at Jan 09, 2014, around 03:30 am in Taxiarchon Park. The calculation of the distance of the two people’s nearest location data is 0.34 km. Since Jan. 09 is a working day, it is important to investigate why the two colleagues have to meet at the early morning instead of seeing each other at company working time. Due to the data absence, the meeting time and location is a best guess we have.

2.2.7, Possible very early meeting of Hennie Osvaldo and Isia Vann at Jan 11 around 3:30am in Spetson Park. Both of them were driving in the time gap. The calculation of the distance of the two people’s nearest location data is 0.01 km. It is important to investigate why the two colleagues have to meet at the early morning instead of seeing each other at day time. Due to the data absence, the meeting time and location is a best guess we have.

2.2.8, Possible very early meeting of Hennie Osvaldo and Minke Mies around Jan 14, 2014, 03:30 am, in Spetson Park. The calculation of the distance of the two people’s nearest location data is 0.xx km. Since Jan. 14 is a working day, it is important to investigate why the two colleagues have to meet at the early morning instead of seeing each other at company working time. Due to the data absence, the meeting time and location is a best guess we have.

2.2.9, Hennie Osvaldo’s has occasional large consumption at Roberts and sons (Jan. 14) and Frydos Autosupply n’ More (Jan. 19). The only large consumption indicates he is treating friends in an occasional gathering. It is important because the occasional gathering happened just before the people missing and is organized by the suspect. But we are not very confident it is related to the people missing since gathering may be used to discuss anything.

2.2.10. Isia Vann has occasional large consumption at Shopper’s Delight (Jan. 10) and Frydos Autosupply n’ More (Jan. 17). The only large consumption indicates he is treating friends in an occasional gathering. It is important because the occasional gathering happened just before the kidnapping and is organized by the suspect. But we are not very confident it is related to the kidnapping since gathering may be used to discuss anything.

2.2.11. Isia Vann has an abnormal pattern at Jan. 19, a weekend day, staying at Frydos Autosupply n’ More in the afternoon without any consumption record. It is important since the date is just before the people missing event and Isia Vann is one potential suspect. But we are not very confident it is related to the people missing since people may wander at weekend.

2.2.12. Minke Mies occasional large consumption at General Grocer at 2014-01-10. The only large consumption indicates he is purchasing many things but it is not a life pattern. It is important because the occasional purchasing happened just before the kidnapping and is performed by the suspect. But we are not very confident it is related to the kidnapping he can buy anything he needs.

MC2.3 – Like most datasets, the data you were provided is imperfect, with possible issues such as missing data, conflicting data, data of varying resolutions, outliers, or other kinds of confusing data. Considering MC2 data is primarily spatiotemporal, describe how you identified and addressed the uncertainties and conflicts inherent in this data to reach your conclusions in questions MC2.1 and MC2.2. Please limit your response to no more than five images and 300 words.

For MC2.1, the GPS data is not completed and discrete in time. If the data are displayed directly on a map as shown in Figure 2.3.1, it will be hard to find useful patterns effectively. Therefore, we draw the pattern by connecting the key points we have. Accordingly, it is easy to see the pattern similarities and find the abnormal pattern as shown in section 2.2.

Figure 2.3.1 GPS data direct display for Isia Vann

In addition, some traffic data is conflict with the normal pattern. For instance, Elsa Orilla seems never go to the office according to the routine pattern. We simply omit the data since the pattern is not reasonable for a normal employee.

Figure 2.3.2 Routine Pattern for Elsa Orilla

For MC2.2, the GPS data is not completed so the meeting place and time of people is uncertain. We calculate the nearest distance to show it is possible that two people were meeting at the given slot.

Figure 2.3.3 Calculation nearest distance example of two people in a given time

In addition, for MC2.2, the consumption data including loyalty card and credit card. The consumption cases includes to use credit card only, to use loyalty card only and to use both credit card and loyalty cards. Therefore, as shown in Figure 2.3.5, the consumption records/usage records for the two cards do not always match. In order to see the consumption pattern completely, we combine the data of loyalty card and credit card in the analysis.

Figure 2.3.4 Loyalty card and credit card usage difference example

Finally, for both MC2.1 and MC 2.2, there are some places has longitude and latitude without any location information. For the important places identified in the abnormal patterns, we mark the place in the generated map with streets, and compared the point to the tourist map to get location information. For example, the red point in figure 2.3.5 is an unknown place. The tourist map showed an apartment nearby. So we guess the apartment is Isia Vann’s home.

Figure 2.3.5 Place mark example of Isia Vann at Jan. 19 afternoon