Hi all,

Thanks for the great work on this. A coauthor and I also scraped the 2015 data for a different project, but the merge w 2021 data you guys have done is particularly useful.

I was wondering if the 'aarakshan' column from the raw pages could be kept in the processing steps so that master.csv retains it for both 2015 and 2021. I know that they're present in the 2015 data [e.g. here - पद का आरक्षण]. In theory, one could reverse-engineer this from the winner's caste variable (if there was perfect compliance and no reserved candidates won in non-reserved cycles, which is a depressing thought), but I found a few values of winner_caste_2021 that I couldn't map cleanly to reservation categories.

I'm interested in studying patterns in incumbency when GPs switch from open to reserved seats (non SC/ST subset of incumbents can't run again) and vice versa (SC/ST incumbents can run again).
Happy to discuss this in private if these are variables of particular interest to you.