ISSS608 2016-17 T1 Assign3 Ong Han Ying - Act 2
| CRIME SCENE DO NOT CROSS....................CRIME SCENE DO NOT CROSS....................CRIME SCENE DO NOT CROSS....................CRIME SCENE DO NOT CROSS |
Plot
| CONTENT |
|---|
|
Part 1 : Who are they? |
Part 1 : Who are they?
Part 2 : When and Where?
Part 3 : Diving Deep: Who are they, again?
Part 4 : What's next?
Highlights
From Part 1 : Who are they?
Question 1: Who are those that have a high communication exchange?
Answer #1 2 Distinct IDS!
A boxplot reveal the following:
For more details, please go to:
Behind the Scene - #Act02 - Fun Fact#01
Question 2: Where and where to they talk to, and received message from?
Answer #2A ID 1278894
The top of the list is ID 1278894, and he/her communication pattern as below;
FINDINGS #1:
- From the heatmap, we can see that ID1278894 sends out a message every 5 min regularly, in equal interval of every 2hour, between 1200 to 2000.
- Further analysis, this ID only send out from entry corridor, and it never moves at all.
- This should be from the server, and likely the application of DinoFun World.
- Look like there are most people came on Sunday.
FINDINGS #2:
- Since the server sent out the message at a regular interval, most of the guest respond within the time frame. But 2 outliers, respond >30min after receiving the message!
- Upon further investigation, both outliers belong to the same ID 1765818. This makes him/ her suspicious.
Answer #2B ID 839736
FINDINGS #3:
- From the heatmap, we can see that ID839736 received messages between 0800 to 2330, there are messages every minute.
- Further analysis, this ID only send out from entry corridor, and it never moves at all.
- This should be from service helpdesk!
FINDINGS #4:
- The helpdesk is receiving the most number of messages between 1200 to 1223.
- It responds readily between 1201 to 1225.
- This high volume of communication is found in the Wetland.
FINDINGS #5:
- The helpdesk is receiving a significant amount of messages between 1440 to 1442.
- This is found in Coaster Alley instead.
From Part 2 : When and Where?
Question 3: When and where are these communication conducted?
Answer #3A An Overview
The distribution of the messages sent over time, excluding server & helpdesk, as below;
FINDINGS #6:
- There is unusual high communication after 2330, on Saturday. Further investigation is required.
Answer #3B Distribution by Time & Venue
The distribution of the messages sent over time by venue, as below;
FINDINGS #7:
- There is a common peak at 11AM at Coaster Alley over 3 days, and 4PM at coaster Alley on Friday & Saturday.
- There is a significant higher spike ar 1700, over at Kiddie Land.
From Part 3 - Diving Deep: Who are they, again?
Question 4: Whom to They Communication With?
Answer #4A Communication between 2330-2335 on Saturday
The network graph selected as below;
For more details, please go to:
Behind the Scene - #Act02 - Fun Fact#02
FINDINGS #8:
- From the graph, we can see that there are 2 main coordinators among those who communicate during this specified timing.
- It might be a tour group that stays late.
- There is also a communication to external. (We shall not make any wild guess first)
- We will take note of this, and find out if anyone that doesn't belong to the group, but still in the park.
Answer #4B High communication frequency at Coaster Alley at 11AM for all 3 days
Outdegree is chosen to display as the node as it is able to display the tour group (or cluster of tourists) easier, as compared to others.
| Friday - Indegree | Friday - Outdegree |
|---|---|
 |
 |
| Saturday - Outdegree | Sunday - Outdegree |
|---|---|
 |
 |
For more details, please go to:
Behind the Scene - #Act02 - Fun Fact#03
FINDINGS #9:
- Based on the out-degree network diagram, we are able to see that people are communication in groups, with certain ID (bigger node), sending out more messages to the others.
- These people are likely to be the leader of the group (or a tour guide).
- Also, we can see small clusters of groups on their own, which might be groups of friends coming together.
- on Friday at 11AM, there is a higher number of message sent (in-degree) to the "external".
- As such, with this high frequency of messages exchange, this is likely to be the showtime of Scott Jones, since it draws a large number of crowd and groups.
- confirming with the movement data later, we will determine at a later stage if these tours can be removed from being a suspect.
Answer #4C High communication frequency at Coaster Alley at 4PM For Fri & Sat only
For communication at 4PM, "out-degree" is selected because it is able to display the group more obviously, as compared to "betweenness" that show distinct group only. This is especially so when there seem to be many small clusters of group communication at this hour.
| Out Degree- Friday | Out Degree- Saturday |
|---|---|
 |
 |
For more details, please go to:
Behind the Scene - #Act02 - Fun Fact#04
FINDINGS #10:
- It was clear that on both days, a big group are attached, and therefore; there was communication among the people.
- The peak of the timing seems to suggest the showtime (either start or end time) of the 6 shows.
- Without a peak on Sunday at a similar timing, it is likely that the "event" was "canceled" or was not being organized since there are supposed to be 6 shows.
Answer #4D High Communication at Kiddie Land, at 5PM
For communication at Kiddie Land at 5PM, Out-degree are analyzed so as to identify the volume of the communication flow sent out, mainly by the "leader" of the group. The network graph as below;
FINDINGS #11:
- The communication is made up of tour groups, too; based on the out-degree diagram.
- However, this is the only timeslot over the 3 days that has a spike, therefore; there might be some special events that occurs.
- This is especially fishy when this area is not near the performance & exhibition, and also; likely after the timing of the event of Scott Jones.
- We are unable to conclude if this were related to the crime, but the sudden increase in the crowd in a place further away from the crime scene, can be fishy.
- In the event that the sudden in the crowd is related to the crime, then; we are able to identify non-suspects -via the tour groups/ communication groups.
- Thus, this is worth to be taken note of, and to be analyzed further with the movement data.
Detective Board
Timeline
