This Node.js project is an exploration of data stream processing. The goal is to exploit a third-party web service and perform analytics on the data stream. The Twitter streaming API is used to perform a basic sentiment analysis on a stream of tweets.
The project uses various other
projects including Express.js, socket.io, Jade,
and jQuery. In addition, the Node.js package Node-Twitter
node-twitter has also been incorporated for easily connecting to the Twitter
stream. Bootstrap and jQuery JavaScript and CSS dependencies should be put in
pulic/javascripts and public/stylesheets respectively. You can put
these files into their respective folders, or you can use symbolic links.
However, the application expects jquery.min.js and bootstrap.min.js
to be in the javascripts folder, as well as bootstrap.min.css and
bootstrap-theme.min.css to be in the stylesheets folder.
The application will connect to the Twitter streaming API and filter tweets with the words 'love' or 'hate'. It will count and determine the percentage of tweets seen which contain these keywords, and it will also display the total number of tweets analyzed. If a tweet contains both words, then that tweet is added to the total count twice, so that the love and hate percentages will add up to 100%.
The project file structure was created using Express.js web framework. The
server code is found in app.js, while the client-side JavaScript code is
in client.socket.io.js. Jade view templates are located in the
views folder. The routes folder defines HTTP routing, but it is
not really used in this project.
The main application is located in app.js, and it requires four environment
variables to be set before running:
TWITTER_CONSUMER_KEYTWITTER_CONSUMER_SECRETTWITTER_ACCESS_TOKENTWITTER_ACCESS_SECRET
These are provided when you register an application for use with the Twitter API.
In addition to the Twitter key requirements, a Node.js environment is also
required to install packages and run the app.js server source, but creating
this environment is beyond the scope of this project.
Once the environment is set up, assuming you are using a terminal in the folder
where app.js is located, you can install Node.js dependencies using:
$ npm install
Then, you can run the application using:
$ node app.js
This will launch the server, and tweets will begin streaming in and printing to
the console. The server will listen for connections on port 3000, and on the
client side, you may navigate to http://localhost:3000 in a web browser.
This will display the sentiment analysis percentages and the total number of tweets
seen since server start-up.
To close the server in a terminal, use CTRL-C.