I’ve been using the real-time Instagram API to look at photos taken in the Mission District of San Francisco.
Based upon some good ideas from my fellow fellows at Stamen, I’ve also started doing some color, sentiment, and keyword analysis on these photos.
I’ve only been collecting the analysis for a week, but here are the most popular photos, keywords, and colors posted in the Mission.
The most popular words are:
“Resolved”, “Like”, “Love”, & “Want”
When I see a new photo in the stream from Instagram, I now do some basic analysis of the post and image file.
The result is a palette of distinct colors in the photograph, including the most dominant color in the image.
Words & Sentiment
Both Twitter and Instagram already include hashtags, but I’m using gramophone to count the keywords in each post. I then perform AFINN sentiment analysis for the keywords. I’m using the sentiment Node.js library by Andrew Sliwinski. At a later stage, I’ll include the corpus of all Mission posts into the analysis. I may use a similar technique to count the most popular emoticons used by Twitter and Instagram users in San Francisco.