Final project for ECON 326. My collaborators on this project (alphabetically) were Bryn Griffiths, Raven Kirkwood, and Aleena Sharma.

• Main Analysis Notebook

• Exploratory Data Analysis Notebook

We study Canadian and American coronavirus data to test two hypotheses:

1. The R_0 for Canada falls within the interval [2.06-2.52]. (This number comes from a Zhang et al analysis on the Diamond Princess outbreak.)

2. Canada "responded better" to the pandemic than the US, since the WHO declared a pandemic. (We quantify this by looking at the reduction in R_0 over the two regimes.)

For (1), we use a simple linear autoregression x_{t+5} = R_0 x_t + e, and examine it afterwords for serial correlation effects (since, e.g., the datapoint (x_0, x_5) and the point (x_5, x_10) are correlated through their shared dependence on x_5).

For (2), we test a hypothesis on the quantity D_canada - D_US, where D_X is the difference in R_0 values for the pre- and post-pandemic regimes for country X.

The data itself is a transformed version of the JHU COVID tracker, frozen April 6th.

The source repo is available here (commit c8d69b5). It's a time-series (JSON format) of relevant information by country and day. For ex:

"Thailand": [
  {
    "date": "2020-1-22",
    "confirmed": 2,
    "deaths": 0,
    "recovered": 0
  },
  {
    "date": "2020-1-23",
    "confirmed": 3,
    "deaths": 0,
    "recovered": 0
  },
  ...

Data is cumulative, and is updated thrice daily from the "official" JHU tallies here.