Earlier this week I published this post where I did some data collection on 201 swipes. That models a population of around 30 000 people, which is small for London. I decided to extend the work and gather 400 swipes of data. This Models over 100 000 people with a confidence interval of 95% and an error margin of 5%. Also, it makes the percentages nice and rounded.

As a reminder, I am 36, straight, cisgender male and my search criteria are Woman, Woman Woman Couples, Non-Binary, Other between the ages of 18 and older (so maximum age). My distance range is set to 400 KM and my profile is physically located in Central London and set to London Core. Lastly, I have the 7 day active filter on. So, without further ado, here are the top line stats. You can see the detailed data and various extracts and charts here.

I am somewhat selective, but not terribly so... #slutlife

Overall, there is a good representation of LGBTQI folks on the app. However, heterosexuals are still a plurality (and maybe a majority if the error margin swings that way). For reference, most sociology studies on the topic report incidence of heterosexuality in the general population at around 90%. Feeld is markedly less straight.

As you can see from the above, there is a big skew in the presented profiles to be around my age or up to 10 years younger. There are relatively few profiles over 10 years younger or 10 years older. I am not sure if this is algorithmically determined or people in those age ranges are more represented on Feeld due to marketing and life circumstances perhaps.

The closest profile presented was 1 KM away, while the farthest was 17 023 KM away. Still, the initial code base for Feeld served profiles by distance first and then by your other criteria. This still seems to be the spirit of the app, as it continues to prioritize distance, though it clearly has other ingredients in the mix as you can see above. Fully 2/3s of profiles are within what I would call my local range - 25KM or less. For reference, that's the average distance between Zone 1 Central London and Zone 6 outer boroughs. For most journeys that is about 45 minutes station to station on the tube (subway).

As part of my swiping, I developed a number of red flag categories. Some are subjective and some are fairly uncontroversial. For example, I label couples behind one profile as Undisclosed Couple, even if the bio mentions it's a couple. I do this because the app has a mechanism for couples to present themselves and doing it in any other way circumvents the search function of the app and ruins everyone's experience. On the flip side, I also label profiles without bios as No Bio, or without a face in the photos as No Face Photo. And thus, we get to the red flag profiles...

As you can see, the majority of profiles have no readily identifiable red flags. Some profiles have multiple red flags, but the really problematic red flags are very rare (fetishizing race, lots of dealbreakers). The most common red flags are No Bio, at almost a quarter of profiles followed by No Face Photo and No Photos (this means they have photos of non human subjects and objects) at a combined 11.5%. Overall, it seems Feeld has done a good job in cleaning up their act with regard to fake profiles as I only encountered 5 over the course of 5 days. Weirdly, Feeld released a tweet/thread saying they have culled 16 000 fake profiles and are implementing further measures to keep them off. So it sounds like that effort has coincided with this endeavor and it shows.

This is all for now. Have a look at the data and let me know if you want any more information. In the sheet, you can also see data about interests and desires. That can't be included here because they are user generated and so there are 567 unique interests for the sample size. Far too many to do a frequency analysis on here.

Edit: Spelling