Chapter 3. Mining Instagram: Computer Vision, Neural Networks, Object Recognition, and Face Detection
In previous chapters we have focused on how to analyze text-based data retrieved from social networks, the structure of the networks themselves, or how strongly people in your network are engaging with the content posted to the platform. Instagram, as a social network, is primarily an image and video-sharing app. It launched in 2010 and quickly gained popularity. The app made it very easy to edit photos and apply various filters. Since it was intended for use on smartphones, it became an easy way to share photos with the world.
Facebook acquired Instagram less than two years after its launch, and as of June 2018 the app had a staggering 1 billion monthly after users, making it one of the most popular social networks in the world.
As social networks have expanded, technology companies have sought new ways of extracting value from all the data being loaded onto their platforms. For example, companies like Facebook and Google have been aggressively hiring people with expertise in the field of machine learning—i.e., teaching computers to recognize patterns in data, to put it very broadly.
The applications of machine learning are myriad: predicting what content you’ll enjoy viewing, which ads you’ll most likely click on, how best to autocorrect the words you’re clumsily typing with your thumbs, etc. As with most things, machine learning can be used nefariously, and so it is important ...