Originally created at The University of Toronto Scarborough's Hack the Valley 2017.
SentiSocial is a Twitter based trend analysis application. This repository contains the SentiSocial backend, which queries the Twitter API for current trends, collects data on each trend and exposes that data via a public API.
Specifically, the backend collects the following information on trends:
- Average sentiment of tweets related to the trend (via the NPM Sentiment Package)
- Popular tweets related to the trend
- News articles related to the trend (via the News Api)
- Keywords found in tweets related to the trend
- Countries in which the trend is trending
- Amount of time the trend has been trending for
- Volume of tweets related to the trend
This data is displayed nicely in the SentiSocial frontend, which can be found here.
The API is documented in api-doc.md.
- NodeJS ^6.9.4
- NPM
- MongoDB
Before running the backend, ensure you have node + NPM installed and run:
npm install
After installation, edit config.js to your liking, making sure to add the
address of your MongoDB server. You will also need to store your Twitter and
News API authentication keys in the following environment variables.
| key | environment variable |
|---|---|
| Twitter consumer key | TWITTER_CONSUMER_KEY |
| Twitter consumer secret | TWITTER_CONSUMER_SECRET |
| Twitter access token | TWITTER_ACCESS_TOKEN |
| Twitter access token secret | TWITTER_ACCESS_TOKEN_SECRET |
| News API key | NEWS_API_KEY |
Then run:
npm start
To spin up the backend.
Jasmine tests are located in the spec directory, to run them, use:
npm test
To run the tests and view test coverage with Istanbul, use:
npm run test-cov
This project uses the JS standard code style.
ESLint is already configured to check the code for adherence with standard. To run it, use:
npm run lint
All code in this repository is licensed under the MIT license. See the LICENSE file for more details. All images and other assets are licensed under CC BY.