[derekjobst]

Dear Tech People is the result of pulling over 70k profiles on LinkedIn across 100 companies and processing all the data with name analyzers/facial recognition tech/and a bunch of actual humans on mturk. The project seeks to answer questions like how many women work at Snap or how many black engineers are at Docker. Hopefully the transparency in this project helps diversity advocates make cases for the change they want to see.

I'm excited to share this project on behalf of the creators. If you're interested in learning more, you can contact them at deartechpeople(at)gmail.com.

[chatmasta]

So you broke LinkedIn terms of service, shared people’s photos without their consent, and asked mturk workers to identify those people’s race and gender?

Something about that really creeps me out. I don’t want someone looking at my photo to determine if I’m Asian or Latino, or male or female or whatever else, then shaming my employer based on the data.

What’s next, auto scanning of resumes to make sure the name sounds like a girl’s name? Requiring photos when applying to jobs to ensure you are “black enough?”

[adinathoughts]

Hi there, I worked on Dear Tech People.

The goal of Dear Tech People is to provide a degree of transparency into diversity data--a topic that frankly a lot of companies feel is taboo and too hard to talk about. Outside the very few companies that have released diversity reports, most companies have poor, incomplete, or no diversity data available internally, and diversity advocates we spoke with found it difficult to make an internal case for inclusion initiatives without data to back up efforts.

Dear Tech People is by no means perfect, but what we've found is that some amount of data, even imperfect data, can give people an idea of where they are and what their goals are. Even some amount of data helps kick off a discussion of leadership or helps people make a case for an internship program that works with HBCUs. Furthermore, the long term vision is to raise the standard of self-reporting, and get companies to fill in incomplete data with complete data that they gather via their own diversity reports. If a company (or even an employee at a company) doesn't think the data looks right, we want to work with them to figure out how to survey their company and get their self-reported data up to standard.

But for now, we work with what we have, and that is public profiles on the internet. You can read more about how we put the methodology together here: http://www.deartechpeople.com/methodology

[overeater]

Are we really now scolding people for crawling the web now, and not following terms of services on Hacker News of all places?

And your stance is that people aren't allowed to judge characteristics about you from your public photo that you put online?

Are you also going to be outraged over people shame conference presenters over a lack of diversity in panels, which requires the same sort of determination?

What's next, requiring filters installed on peoples' eyes to prevent them from seeing race/gender in photos?