A still-hypothetical project to build a pipeline for aggregating political memes from Facebook.
¯\_(ツ)_/¯
- Text acquisition: Luckily these things appear to be fairly OCR-able
- Face (ID) classification: Spot the face, tag the face. I think there's good coverage for politicians in many face training DBS. If so we'll have to aggregate.
- Its an infrastructure: let this run as a service that's easy to set up and use as part of another project
- Its really-well documented: Do you ust want "All The Facebook Memes" on your computer? Great - thats straightforward. Do you want to use this for an in-depth image and content analysis pipeline? Also incredibly straightfoward.
- Can run as a hosted service for others
- Doesn't violate anyone's ToS / IP rights