1 min readMay 4, 2020
It is an interesting taken and thanks for sharing. However, I have a few remarks (related to possible biases):
- What is the % of profile pictures? (It can vary by language.)
- I wouldn’t be surprised if there % of these change a lot by age and gender group. Did you try to compare data with gender inferred from the first name? (There are a few APIs for that; also ‘he’ may be not the best word for an unknown programmer.)
- For age I really suggest using violin plots or (my favourite!) swarm plots.
- I for the comment data, I would love to see a scatter plots with the ratio of positive, and the ratio of negative comments. (Also, curious how does it compare to yearly StackOverflow survey, in which they ask about loved and hated technologies.)