Hello, As I mentioned in the diversity ad-hoc meeting this week I have been tweaking the tool I use to estimate gender of RIPE meeting attendees. I finished these tweaks and have some (hopefully) corrected results. Just to re-cap, the idea is to look at the first name of the people on the attendee list and guess if they are a woman or man based on that. For example "Martina" is probably a woman and "Tim" is probably a man. The tweaks were: * Not trying to guess the gender of single-letter first names, like "A" or "P." (I assume these are abbreviations and not actually people named "N" or the like, and we do not have enough information to make any guesses). * Splitting off the first name of a compound name. So while the name API doesn't know that Thor-Henrik is most likely a man, it does have an opinion about Thor. * Going through the names that the API does not know and see if we can figure out the gender based on other things. This includes looking up information online (for example in directories of baby names) and also recording names of people that I have met (Rudiger, Maldwyn, Denesh, Olafur, Iratxze, and so on). The overall picture doesn't change too much: on average we have 0.25% fewer women at each RIPE meeting, although in more recent meetings the correction is in the other direction. I have updated with the latest information from the RIPE 76 meeting page, and attached the "corrected" versions of: 1. A graph showing percentage of men & women at each meeting 2. A graph showing just the percentage of women at each meeting, along with a trend line 3. A graph with the percentage of women at each meeting, with the trend line extrapolated to the magical day where 50% of attendees are women 4. My scratch spreadsheet for all of this I've also updated the GitHub: https://github.com/shane-kerr/ripemtggender Now I'm late to the RIPE dinner, but totally worth it. 😉 Cheers, -- Shane
