This is a stretch goal! If a good method is determined for identifying TravisCI usage, then we should explore if there is any correlation between that and the event type. We are currently assuming that Watch events will provide the highest frequency here but that might not be true. We should see if other event types produce repo samples with better results.
This is a stretch goal! If a good method is determined for identifying TravisCI usage, then we should explore if there is any correlation between that and the event type. We are currently assuming that Watch events will provide the highest frequency here but that might not be true. We should see if other event types produce repo samples with better results.