Glimpse

A prototype tool to give users a "glimpse" of attention on a platform based on a query with keywords and start/end dates. Created to use in teaching as a temporary and exploratory solution. Right now this supports querying for the amount of attention on a number of platforms:

• Online news (via Media Cloud)
• Twitter (via their academic API track)
• YouTube (via their API)
• Reddit submissions (via Pushshift.io)

Speculative Roadmap

This was a weekend project created to support teaching in a Digital Storytelling and Social Media class. Going forward, it would make more sense to:

• implement a uniform "simple query" mode, so first-time users don't need to learn the query syntax specific to each platform
• try to determine a denominator so we can show percentage of attention over time for each platform in one chart
• support filtering content by platform (ie. channels on Reddit, collections of media sources for online news, etc)
• add support for more platforms via their APIs

Installing

This requires installing redis (for caching results). Then:

pip3 install -r requirements.txt

Running

You can run this locally for development by creating and filling in a .env file based on the .env.template.

Running in Docker

docker image build -t glimpse .
docker container run --rm -it -p 8000:8000 -e MC_API_KEY=keykeykey -e CACHE_REDIS_URL=urlurlurl -e TWITTER_API_BEARER_TOKEN=tokentoken -e YOUTUBE_API_KEY=keykey glimpse

Deploying to Dokku

1. dokku apps:create glimpse
2. dokku redis:create glimpse-cache
3. dokku redis:link glimpse-cache glimpse
4. dokku config:set glimpse MEDIA_CLOUD_API_KEY=keykeykey CACHE_REDIS_URL=urlurlurl TWITTER_API_BEARER_TOKEN=tokentoken YOUTUBE_API_KEY=keykey
5. dokku config:set --no-restart glimpse DOKKU_LETSENCRYPT_EMAIL=emailaddress
6. dokku ps:scale glimpse web=6

Credits

Created by Rahul Bhargava, Assistant Professor of Journalism and Art + Design at Northeastern University.

• Data Culture Group
• Media Cloud