Geographic Twitter visualisation server

Configuration

Five environment variables must be set:

SENTIMENT_SERVER_HOSTNAME

This is the hostname or IP address of a server running the code that can be found at this repo. This is required for both visualisations.

TW_KEY
TW_SECRET

These are required for the /live endpoint. An app needs to be set up in the Twitter developer portal and the credentials set to these variables.

GRAISEARCH_USERNAME
GRAISEARCH_PASSWORD

These are required for the /scenario endpoint, they must be valid credentials for read access for the dataset on the server graisearch.scss.tcd.ie

Deployment

The app is capable of deployment on Heroku's cedar stack with a git push.

Alternatively the app requires rubygems to be installed. You can then run bundle install from the root directory and rubygems will install all dependancies. The app can then be launced with rackup -p :port_number

Features

This app uses Mapbox, Audiolet, Twitter and the Stanford CoreNLP Toolkit to visualise and synthesize music to tweets and their sentiment over two settings. Live over San Fransisco and over the course of the Dublin Marathon in 2014.