This repo contains a county-level Kinsa US Health Weather Map data scraper.
Kinsa is a company that makes Internet-connected thermometers and publishes the subsequent temperature readings in a "US Health Weather Map" Such readings can be classified as an illness when the user has a fever, and from there Kinsa uses an epidemiological model to estimate observed %-illness vs what would be expected under their model.
The scraper consists of a Jupyter notebook found in ./notebooks/scrape_kinsa_health_weather.ipynb
1. Python requirements
Make sure you have installed Jupyter notebook
Now install the required Python libraries:
pip install pandas matplotlib ipywidgets seaborn selenium
2. JS resources to enable IPython widgets
- Install node, the Javascript framework
- Run the following commands:
jupyter nbextension enable --py widgetsnbextension
jupyter labextension install @jupyter-widgets/jupyterlab-manager
jupyter notebook ./notebooks/scrape_kinsa_health_weather.ipynb
And then follow the instructions therein. The scraped data will be saved to .csv files in ./data
with the following columns:
Column | Description |
---|---|
condition | "observed" %-illness or "atypical" %-illness |
date | date of %-illness reading |
county | county name |
state | state abbreviation |
fips | 5-digit (county, state) FIPS code |
illness | %-illness recorded for county |