ISSS608 2017-18 T1 Assign Fam Guo Teng Origin and Epidemic Spread
|
|
|
|
|
|
Contents
Visualizing the Data - The Origin of the Epidemic
Using the reported symptoms of the mysterious outbreak, the various microblog messages were labelled with the identified symptom keywords for plotting. Some of the keywords used can be divided into two main groups: Respiratory-related symptoms (for e.g. cough, breath) and Gastrointestinal-related symptoms (for e.g. diahrrea, nausea, vomit).
When plotted on the map of Smartpolis using the given Latitude and Longitude information, the microblog data points revealed some interesting patterns. Most of the respiratory-related data points were clustered in the Downtown area, and most of the gastrointestinal-related data points were amassed around the south-western river banks of Vast River.
Figure 1 highlights the ground zero location of the epidemic in a yellow circle, this is the location of a truck accident on the Interstate-610 bridge (I-610). The green outline shows the data with respiratory-related symptoms, possibly caused by a westward breeze. This is in-line with the behaviour of an airborne transmission. The red outline denotes the data describing gastrointestinal-related symptoms, in-line with the behaviour of a waterborne disease transmission. The blue circles display the various hospitals in Smartpolis that have an influx of patients during the epidemic.
Keyword Analysis
Furthermore, doing an investigation of the occurrence of the keywords sorted by date (Figure 2) revealed that there was a sizable increase in the reporting of the symptoms some time after the date 05/17/2011.
From this, one can infer that the source of the symptoms is quite likely to have appeared on the date of 05/17/2011 and it is worth investigating.
Text Analysis
Performing text analytics on the microblog messages of the events on 05/17/2011 reveal several notable events that happened (Figure 3).
Using both the scatterplot visualization, SAS JMP Pro 13 Text Explorer, and the keyword-per-date information before these symptoms occurred, one can see that the data has clues pointing to the truck accident that occurred near the I-610 bridge location. This happened on 05/17/2011 at approximately 11:00 AM and the accident caused a fire during the incident. The truck accident lies right in-between both infected regions, making it a likely cause of the epidemic.
Weather Analysis
| Date | Weather | Average Wind Speed | Wind Direction |
|---|---|---|---|
| 05/17/2011 | Clear | 8 | W |
| 05/18/2011 | Clear | 9 | W |
| 05/19/2011 | Cloudy | 9 | WNW |
| 05/20/2011 | Cloudy | 5 | NW |
Since there was a breeze blowing from the West during the day of the accident, and for the following few days of the outbreak, it is possible that the respiratory symptoms in the Downtown area were caused by fumes from the truck fire that had ignited after the accident.
Stated in the microblogs, the truck had also been spotted spilling cargo during the incident. The cargo spill could have somehow contaminated the Vast River with the water flowing south at a steady rate of three miles per hour. This could have also infected anyone downstream who had used the river for water-based sports and recreation, causing them to display gastrointestinal symptoms.
