Alone Together

Demo of the friendship paradox via Twitter social network. The friendship paradox states "most people have fewer friends than their friends do." This is true for most social graphs because popular people will push more people below the average.

Instructions

Prerequisites

1. A Firebase account with billing enabled. Billing must be enabled to call third party APIs like twitter.com.
2. The latest copy of the firebase-tools.
3. A Twitter app. You can create one at apps.twitter.com.

Setup

1. Clone this repo.
2. Set firebase-tools to use your project with firebase use <YOUR_PROJECT>
3. Use firebase functions:config:set to set twitter.consumer_secret, twitter.consumer_key, twitter.access_token_secret, and twitter.access_token_key. You can get these values from apps.twitter.com
4. Deploy with firebase deploy

Usage

1. firebase deploy will print out the URL of your http function (it will end with "/rating")
2. Go to this URL. It expects the query parameter "?username="

Testing locally

1. Make a local copy of your runtime config by running the following command from your functions directory: firebase functions:config:get > .runtimeconfig.json (We highly recommend you include this file in your .gitignore)
2. Start the functions shell with firebase experimental:functions:shell or create a local HTTP server for the rating function with firebase serve --only functions

Corners cut

This is just a tech demo so a number of corners were cut to save time or make code more readable:

1. This program will only analyze a user's first 5000 followers. Support for more followers requires pagination support in the getFollowers helper method.
2. This program hard codes all API calls to be made against a single user. A better option would be to use application-level credentials. This requires a manual OAuth call with the twitter node module. It comes automatically in the twit node module, but this SDK has a much more verbose return type and makes the rest of the code less readable. The best option would be to accept Twitter OAuth tokens in the rating function's Authentication header and use that for Twitter API calls.