This week, I decided to experiment with a few different ideas and technologies in order to further develop my thesis project. Here are some of the projects/experiments I worked on:
Experiment #1: Chrome Extension (for Facebook).
A chrome extension that swaps all the pictures in your Facebook feed with the logos of the advertisers that currently have your contact information.
I started by downloading my entire Facebook archive (do it yourself). I found a list of all the advertisers who have my contact information from Facebook – more than 200 entities in total. This information shocked me, especially because a number of them were data collection companies and politicians running for office.
I took that list of advertisers and decided to scrape Google Images to download all their logos.
Then I shifted gears and built a Google Chrome extension that swaps all the images on Facebook for any images of your choosing. I wrote code that picks a random image from the folder of advertisers every time the page reloads.
Personally, I found that the advertisements for various Senators and politicians to be the most intrusive and unwanted.
Experiment #2: Facemash (for Facebook and LinkedIn).
A Python script that scrapes tagged images from Facebook and LinkedIn, and then identifies & overlays the faces using OpenCV.
I wrote several different Python scripts. One scrapes all the Facebook images you (or your friend) are tagged in. Once you have those images, you can run another script to identify the face and then overlay the faces on top of each other.
I used a few different Python packages and models, including OpenCV for facial recognition/warping and dlib for overlapping the images. Read detailed instructions here (many thanks to Leon for his helpful workshop).
Here are some examples for me and my sisters:
I wrote another Python script that scrapes all the profile pictures from your LinkedIn connections. I was able to scrape the first 40 connections and then ran those images through the facemash script. This is what my average LinkedIn connection looks like:
Experiment #3: Aristotle Search (for Twitter).
I’ve been thinking a lot lately about how we search for and filter information online, and the ways in which Twitter and Google, for instance, make decisions for you about what’s most relevant. What if you wanted the ability to filter Twitter results according to a different set of criteria?
Inspired by Ted Hunt’s Socratic Search, I built a sister search engine called Aristotle Search that filters Twitter results according to Aristotle’s criteria for persuasive argument: logos (appeal to logic), pathos (appeal to emotion), and ethos (appeal to ethics).
The search engine is meant to be an exercise in speculative design that allows us to think about how a redesign of social platforms would change how we approach and engage with them. What if you approached Facebook with the intention to strengthen your relationship with family or reconnect with high school friends? What if you approached Google with a desire to challenge your own assumptions or seek clarity? (see: Socratic Search)