Brand Sentiment Analysis Extension for Twitter

Are you a Twitter enthusiast who wants to quickly see the sentiment behind the tweets on your timeline? Or maybe you’re an organization and want to understand the sentiment towards your brand on Twitter. This browser extension is for you. It performs sentiment analysis on tweets and displays the results in an easy-to-consume format. This project provides a hands-on opportunity to learn about machine learning in practice, as well as how to create browser extensions.

Features

• Real-time sentiment analysis of tweets
• Visual representation of the sentiment of tweets in a pie chart

Requirements/Prerequisites

• Docker should be installed on your local machine
• An entitlement key from IBM, which can be obtained here

Installation

Starting the Sentiment Analysis Model locally

If you start the sentiment analysis model locally then you must ensure you have the IBM entitlement key from the requirements/prerequisistes section as well as have docker installed. Then run the following commands in your terminal in this directory:

IBM_ENTITLEMENT_KEY=<your-key>
echo $IBM_ENTITLEMENT_KEY | docker login -u cp --password-stdin cp.icr.io
docker build -t nlp-model .
docker run -d -e ACCEPT_LICENSE=true -p 8080:8080 nlp-model

This starts up a container with the Watson Sentiment Analysis model on port 8080. You can then use the following value for your container URL in the Javascript files.

const containerUrl = "http://localhost:8080/v1/watson.runtime/nlp.v1/NlpService/SentimentPredict";

If you deploy your sentiment analysis model somewhere else, be sure to adjust the containerUrl.

FireFox

1. Clone or download the repository to your local machine
2. Open a new FireFox window
3. Navigate to about:debugging#/runtime/this-firefox
4. Click on Load Temporary Add-on... and then find the manifest.json file from the cloned repository

Chrome

1. Clone or download the repository to your local machine
2. Change all references of browser to chrome

To use the extension in Chrome, all instances of browser must be changed to chrome. These instances are in the Javascript (.js) files. So, be sure to double check the popup.js, tweet-sentiment.js, and background.js files.

3. Open a new Chrome browser window
4. Navigate to chrome://extensions/ and ensure the Developer mode setting is on
5. Click on load unpacked and select the root folder of this cloned repository

Usage

1. Open Twitter in your browser
2. Click on the extension icon
3. Search for a topic or brand of your interest (e.g. Spotify).
4. Scroll through the feed and open the popup to see sentiment analysis results through the updating pie chart

Note

• The extension is currently only compatible with twitter.com but this project can be used as a template to branch to other social medias

Example

This is an example of the browser extension in action. In the image, Spotify was searched, and the data was populated by scrolling down the timeline for a little bit. We can see Spotify has more of a positive sentiment meaning people are tweeting good things around Spotify's brand. The data is displayed in a pie chart and can be shown or closed by clicking the browser extension popup button. As the user scrolls the data populates more. If the user would like to move to a new query, the reset button can be clicked to reset the current results.

License

This project is licensed under the MIT License. But in reality, use it as you please.

Contact

If you have any questions or feedback, feel free to open a ticket.