Fun and informal data analysis projects.
We use a GLM approach to find covariates that have a high influence on the odds of lethality in a car accident. We use a dataset of roughly 250000 recorded car accidents in the UK.
Methodology and data
Live diagram of a Covid19 outbreak on Reddit started from this post. By commenting in this thread, you potentially infect redditors that you have interacted in the past -- watch it spread!
Every user commenting in a monitored thread gets their recent commenting history scanned and may infect users whose posts/comments they have commented. Once we have infected so many people in a subreddit that it is computationally not feasible to keep spreading on a case by case basis, a Covid SIR model takes over and simulates further spread in that subreddit.
We interact with reddit via PRAW and the fronend is build with vue.js and c3.js for the charts.