COVID-19 Spread Analysis for an NGO
Context
Early in the pandemic I did a short pro bono project for a Dutch NGO. The question was whether Facebook mobility data could say anything useful about how COVID-19 was spreading, and there was a second thread looking at the text of online help requests.
What I did
- Analysed Facebook mobility data against the spread of COVID-19
- Ran text analysis on incoming online help requests
Results
The honest result on the mobility side was negative: the data wasn’t good enough to predict how the disease spread, and it was worth saying so clearly rather than dressing it up. The text analysis was more useful, and helped the communications team improve their help forms.