Data source: https://covidtracking.com/
High-level Hypothesis statement: "Higher positive test rate and/or low numbers of tests would imply a faster rate of growth later in the positive cases curve."
This hypothesis can be broken down into Hypothesis A and Hypothesis B, below.
Hypothesis A: "Higher positive test rate implies faster rate of growth later in the positive cases curve"
hypo_a.ipynb: Currently, Tea does not support modeling (working on providing this soon!), so I tested a simpler hypothesis: Higher positive test counts imply higher growth rate (as measured by increase in positive tests from yesterday, which is a metric reported in the data).
To be totally honest, I'm not sure this is a totally accurate operationalization of the original hypothesis, even without modeling capacity.
Hypothesis B: "Low numbers of tests would imply a faster rate of growth later in the positive cases curve"
hypo_b.ipynb: Very similar to Hypothesis A, above. The main difference is that total number of tests, instead of positive test cases only, are considered.
Hypthesis C: Higher testing rates is positively related to higher death count.
Collaborators: Ben Zorn, Emery Berger