Data Visualization of the Spread of a Virus on a Campus

From CSclasswiki
Jump to: navigation, search

--Thiebaut 15:55, 26 January 2015 (EST)

Lujun Jian's Independent Study Page


In this research we study how a disease such as Ebola, or the Measles, would spread on a closed campus such as Smith College, with its 2400 students. We have picked the Measles, as it spreads more easily than Ebola. The main engine for this project is a discrete-even simulation program that runs several hundreds of times, always starting with a random student being infected by a virus, and starting taking classes at the beginning of the semester. We have generated random, synthetic weekly schedules for all the students at Smith. A random schedule consists of the meeting times of 5 randomly picked courses and lab(s), meeting in different academic buildings on campus. We do not attempt to limit the enrollment in any given class. The schedules for all students are kept in a mySQL database. When an infected student enters a classroom, every other student in the room gets infected with a probability p extracted from the susceptible, infected, recovered (SIR) model for the given virus (typically ~1%). A student changing state in the SIR model, for example going from susceptible to infected generates new events in the discrete-event queue. The discrete event-simulations run until the queue is empty, or until the whole population is infected, or until the end of the semester (14 weeks). The statistics for each run is kept in the database, and allows the generation of a heat map of the average spread of the virus in campus dorms as a function of time, or a plot of the growth of the infected population of students as a function of time, for different parameters relating to the virus.


This poster was presented at CCSC-NE 2015, at the College of the Holly Cross, Worcester, MA.


Coding and the Logic of the Code

Go to this page to explore the code of this special study.

Log of Activities

1/26/15 Meeting

  • Define plan for this semester
    • Phase 1: research, state of the art.
    • Phase 2: review what particular questions need to be answered, and generate plan of attack
    • Phase 3: coding + data visualization
  • Define schedule (can be refined as semester progresses)
  • Identify faculty resources
    • R. Dorit
    • J. Carris
  • Come up with list of questions for faculty resources
  • Organize and maintain this wiki throughout the process.

2/5/15 Meeting with Prof. Dorit and Jon Caris

  • Disease we want to focus : Ebola/Measles
  • Parameters of disease should be concerned
  • Map tools
    • one card data
    • Google API - heat map
  • Valuable info we should keep and generate - path of individual spread track -- network about infected individuals

2/17/15 Meeting with Prof. Baumer

  • Figure out what Plot would be most useful plot or graph of disease spread
  • KML - generating heat map
  • Density Plot of number of stimulations versus percentage of infected students
  • Look into SIR model

Plan of Study for This Semester

--Ljian 10:33, 27 January 2015 (EST)

  • Look in to how Ebola or any virus spread in the campus
  • Timeline for research: Identify problem --> Identify 1 or 2 problems to solve --> Coding --> Simulation

  • Research
    • search published researches (get help from librarians)
    • talk to human resources (faculty)
      • Prof. Robert Dorit
        • Identify parameters
          • how long contact between 2 people for passing virus
          • distance minimal for passing virus (e.g. sitting in a same classroom for an hour, walking to each other, being roommates)
      • Jon Caris (GIS Specialist)
        • tools for mapping dynamically area on a map (heat map- growth of area over map, GIS info for building (GPS)
        • Traces
          • Apps for generating traces? (Google?)
          • Format of the traces?
            • data representation(DPS position + time)
      • D.T. (programming)
        • mobile app. (QT5 - iphone)
        • data visualization
        • data base SQL
        • simulate students(synthetic traces)
          • All houses --> GPS position of all
          • All students distributions/houses
          • catalog --> information about location of the classrooms / class schedule
          • model--> pick a student in a given houses. At random, pick 4 classes from catalog that do not intersect in time. Generate a trace from house - class 1 - class 2 - lunch at random dinning hall/ the one closed to class 2 - class 3 - class 4 - dinner - random evening destination
        • graphs

Discoveries from Reading Literature

Gallery of Interesting Charts/Data Visualization

Go to this page to see a gallery of interesting charts discovered in the literature.


Pandemic: Web game about virus spread

Papers describing research close to our project

Papers about Measles

Papers about Ebola