Analyze profile photos to glimpse who works where.

In Your Face is an attempt to use public data to analyze diversity within the workplace. In particular, the social network LinkedIn includes a huge amount of self-published data about who works for which employer. LinkedIn does not publish demographic information such as gender, age, or ethnicity. However, we can get a glimpse of that information by analyzing profile photos. We use the free Face++ API for this purpose.

Imagine you wanted to tally up how many men vs. how many women were on LinkedIn for a particular company. You could search on the company, then go page-by-page through the results, counting up the number of profiles with a man's face compared to the number of profiles with a woman's face. In Your Face tries to automate that process.

In Your Face is a browser extension for Google Chrome. It would be more practical to provide users a website, but LinkedIn does not provide a search function via their API. A browser extension allows us to examine the photos for any given search. The extensions does not store the data at any time, but merely sends off publicly-accessible image URLs to the face analysis API and tallies the results.

It's important to note that the face analysis API often makes mistakes on individual photos. However, we think that for large enough searches, the results should be representative.

Go to LinkedIn and search for any company. Look for a bunch of small profile photos on the right, and click on the "See All" link. You can make additional filters on the right.

Alternatively, you can go directly to the search page and then filter the results. This way you can also search within your own network.

Once you have your search done, press the In Your Face icon and the "Start Analysis" button. You should see the photos and analysis results go by. You can press "Stop" at any point.

1. Clone the repository.
2. Install dependencies (bower install).
3. Sign up for a Face++ API account.
4. Copy the config.sample.js file to config.js. Edit the file to copy in the API key and secret from Face++. You can also change which of charts are included by editing the CHARTS parameter. The complete list is ["gender", "race", "age", "glasses", "smiling"].
5. Load the unpacked extension into Chrome

If you make changes to the code, don't forget to reload the extension (in the extensions tab) as well as reload the LinkedIn page that you are searching on.

Requires ClojureScript.

To get an interactive development environment run:

lein figwheel

and open your browser at localhost:3449.

To clean all compiled files:

lein clean

To create a production build run:

lein do clean, cljsbuild once min

And open your browser in resources/public/index.html. You will not get live reloading, nor a REPL.

Use a web server such as nginx to redirect requests to website/public/resources/index.html. An example nginx config file is included in this repository.

We use the great library C3.js to make the pretty charts, and the free Face++ API for face analysis.

The developers of this project do not have any connection with LinkedIn. Neither the developers nor the project stores any profile information coming from LinkedIn. All use of this extension is the full responsibility of the user.