tweet_codes

What is this?

This is a minimalistic web-app that can be used to provide tweets (saved in a MongoDB) with codes based on your own coding-scheme, in order to analyse the tweets later throught QDA, machine learning, etc. The tool was developed for a ML-project in social sciences, where i wanted to train the readme2-Algorithm in R to automatically code huge amounts of tweets in order to investigate how certain attitudes towards a topic/hashtag change over time within the Twitter community.

How is it built?

It is built as a full stack MEVN-app based on MEVN-CLI thought for local use with a local MongoDB installation.

The JSON-model used in the app to map the tweets is based on the collections generated by TweetScraper, which I use to scrape tweets and save them directly into a MongoDB collection.

How to use it?

1. Install node.js, npm, MongoDB and MEVN-CLI.
2. Get TweetScraper and edit settings.py to save the tweets into a MongoDB collection. Scrape some tweets then.
3. Clone this repository: git clone https://github.com/nomaad/tweet_codes.git
4. Edit server/config/config.js to point to your local MongoDB collection
5. To add your coding scheme reference, edit client/src/components/CodeScheme.vue
6. Run mevn serveand start the server first, then open another terminal window and run it again to start the client. Happy coding.
7. Analyse your coded datasets in R/Python/your favored data science tool.