ISSS608 2017-18 T1 Assign RACHEL TONG ConcludingRemarks

Home

Preface

Origins

Transmission

Containment

Coda

Coda

To wrap up the analysis, some final precautions to consider include the following

Containing the Airborne Disease

An important aspect for consideration is that the flu-like illness displays strong evidence of being airborne and having lasting effects over multiple days (from the spill on the 17th, with effects lasting even up until the morning of the 20th). As such, the weakening of wind on the 20th means that to insure containment of this illness, the cargo spillage area on highway 610 and affected surroundings should be properly decontaminated. Appropriate agencies should also oversee that residences and commercial buildings guard against potential contamination, and workers and residents should be educated on precautions to minimize exposure to the contamination.

Managing the Effects of Water Contamination

A possible precaution to take to prevent the exacerbation of the situation could be to advise the people of Smartpolis, especially inhabitants of the affected areas, to avoid activities in or near the Vast River for a quarantine period. Officials could issue a public explanation about how evidence indicates the likely contamination of the river and describe the potential dangers of the contamination to public health. In addition, with the water-borne nature of this illness, city officials should strongly consider immediate testing and further monitoring of the water supplied to households and firms from local distribution systems to assess whether this contamination has had an impact on the water supply.

---

Appendix

This section includes additional links to interactive plots developed, and general description on the data preparation and analysis methodology.

Web-based Interactive Plots

• Changes in Population (Daytime vs Baseline) by Region
• Microblog Message Trends by Type
• Microblog Bubbleplot (Tableau Version)
• Wind Direction Plots

Data Preparation and Analysis Methodology

The major parts of data preparation and analysis revolved around the following

• Microblog Messages file
  • Cleaning and formatting the dates and times; an important step which involved aggregating timestamps to the nearest 15-minute interval
  • Identifying and tagging major events: This was carried out iteratively using a combination of phrased-based tagging, selective filtering, and the text analysis feature of JMP Pro. To begin with, a new column ('Tags') was first created to tag messages related to the symptoms. This was carried out using the row select function (Rows > Row Selection > Select Where...) via symptoms provided in the provided information (e.g. Nausea, shortness of breath, cough, etc.). The row select function was also iteratively used to sieve out potentially helpful messages or events that may shed light onto the disease origins or transmission modes. For example, in line with the initial theory about congregations of people facilitating person-to-person transmission modes, a selection of messages by the word 'game' followed by applying the text analysis feature (Analyze > Text Explorer) revealed that basketball was the most popularly mentioned sport, followed by baseball, and subsequently football. In this manner, searching for phrases like 'explosion' (factory or plant explosion) and 'spill' (chemical spill) revealed the occurrence of major events like the Smogtown explosion and truck cargo spillage along highway 610. Further, to minimize noise, some negative filtering was also carried out (e.g. by selecting messages with spill that did not contain phrases about the BP gulf oil spill; which is a global event and unlikely to affect Smogtown in this context).

• Weather Data
  • Beyond the static calendar plots shown in the previous pages, a bubbleplot was also created on the wind data for comparison with the message-based bubble plot. This was carried out by first obtaining a list of unique timepoints equivalent to the range of timestamps of the messages, followed by merging the associated wind data with the relevant dates. Subsequently, coordinates were derived to plot the destination direction of wind (a more visually-intuitive aid relative to the source direction, as was provided in the original file).

Home

Preface

Origins

Transmission

Containment

Coda