A neat tool for Twitter got some renewed attention this week, reigniting conversation around the problem of subconscious gender bias and the proclivity of men to follow other men on social media. It is, essentially, one way to help diversify your feed to make it less of an echo chamber and something more like what I imagine CEO Jack Dorsey wants it to be (and which, dear reader, it definitely is not).
Proporti.onl, an open-source app by engineer A. Jesse Jiryu Davis, is a neat tool to help you analyze the estimated gender distribution of your followers and the people you follow on Twitter, be they male, female, or gender non-binary. It isn’t new—Davis wrote a blog post about it back in 2016. But it does continue to crop up every few months—in reports at the New York Times and the Forward last year, as two examples, and at The Next Web this week—and is evergreen in its functionality for surfacing a pervasive problem on the hell site that is Twitter.
First, a little bit of background. Davis wrote in 2016 he realized that he, like many men, mostly followed and was followed by other men on Twitter. Davis also realized he wasn’t followed by any non-binary individuals that he knew of. He wrote that when he used Twitter’s own analytics tool to pull up his follower ratio, non-binary individuals were—and, in fact, still are—excluded from the demographics section.
“Back in 2016, I read that Elon Musk, Tim Cook, and other famous men in tech hardly followed any women on Twitter,” Davis told Gizmodo by email. “I was anxious to know whether I followed a representative group of people on Twitter, and I was curious to see how hard it would be to estimate. I was also curious to see if anyone else would think it was important to know the gender distribution of their Twitter friends, so I put the tool online.”
The tool isn’t close to perfect. It relies primarily on the pronouns people offer in their Twitter bios (e.g. she/her, they/them) to identify the gender identity of the Twitter contact. Where there isn’t one provided, the program assumes an individual is a man or woman based on how common that name is for either gender. Where the program can’t tell—e.g. with a name like Pat, as Davis noted on Twitter in December, which is common among both men and women—or where the Twitter account is an organization, the tool won’t categorize it. It’s because the program is always guessing and often wrong that Davis says it has also faced some criticism.
“Being misgendered is very painful for most people, even on proporti.onl where the misgendering is hidden within an overall statistic. I had a difficult conversation with someone who thought I should never have built the tool,” he told Gizmodo. “I think that pain is worth it, in exchange for encouraging Twitter users to follow more women and nonbinary people, but I’m not sure.”
Subconscious gender bias is hardly the lone reason that Twitter can be a toxic expanse of irrepressible despair, to be sure. The last thing anybody needs is an isolated echo chamber of confirmation bias, narrowing their ability to observe, listen, and learn from experiences different from their own.
But as Davis noted, there are likewise many important conversations being had on Twitter, and like other social media platforms, it offers a space for communities to easily connect and communicate. Davis noted that he personally is influenced by the opinions of people he follows on Twitter, as many of us are.
“Even more than when I first built proporti.onl, I want my opinions to be influenced by a diverse group of people,” he said.