Personality Insights

Social Media Analysis

The Personality Insights application demonstrates how you can use the services Insights for Twitter, dashDB, and Personality Insights to gather and analyze Twitter data.

Introduction

This Personality Insights sample application has been created so you can deploy it into your personal DevOps space after signing up for Bluemix and DevOps Services. You will attach the Insights for Twitter, dashDB, and Personality Insights services. Once the application is set up, you will be able to search for tweets matching certain criteria, load them into a database, and analyze the personality of any Twitter user for which you have enough data in the database.

If you already know how Bluemix works and you want to automate forking the project, and editing the launch configuration, click the Deploy to Bluemix button below. Once the deployment is finished, jump to How the app works.

Fork the Project to your personal DevOps space

Our source code is stored on GitHub. To fork the project into your own repository navigate to https://github.com/cfsworkload/personality-insights.

1. Navigate to https://hub.jazz.net/.
2. Click CREATE PROJECT.
3. Select Link to an existing GitHub repository.
4. Select Link to a Git Repo on GitHub.
5. Choose your newly forked project from the dropdown menu that appears.
6. Choose your Region, Organization, and Space. Generally the defaults will be sufficient.
7. Click CREATE. This will fork your GitHub project into IBM DevOps services, and redirect you to your new project.

Edit launch configurations

Next, you'll edit the launch configurations in order to deploy your app.

1. After the project is successfully forked, click EDIT CODE in the upper-right corner of the screen.
2. In the top navigation bar, click the drop-down menu and click the pencil icon to the right of the app name to edit the launch configuration. A dialog box will appear and you will be required to enter information about where the code will be deploy.
3. Check that your Target, Organization, and Space are correct.
4. Enter a unique name in the Application Name field. This creates the route that you will use to navigate to your web app after deployment.
5. Enter the same application name into the Host field.
6. Verify that the Domain field is correct and click Save.
7. Click the play icon to the right of the drop-down menu to deploy your application. This deploys the application to Bluemix with all of the necessary services.

How the app works

1. Go to the app's web interface by clicking the Open the Deployed App button in DevOps Services or by clicking the Open URL button in the Bluemix dashboard.
2. At the top of the page there is a query builder you can use to search for Twitter data.
3. The Twitter Count button returns the number of tweets that match the given query. The Twitter Search button returns 20 tweets that match the query and loads them into dashDB. (You can change the number of tweets returned in the code.)
4. Once you have enough information from a single Twitter user (usually 20 tweets are enough), you can enter their Twitter username in the Personality Insights section.
5. Clicking Analyze Personality will send all of the user's tweets to the Personality Insights service, and if there's enough data it will display a textual summary and visualization of the personality data the service returned. If an error is displayed, there may not be enough data from that Twitter user to perform an analysis.

Search for Twitter data through Insights for Twitter

The Insights for Twitter service offers two plans - Free and Entry. The Free plan, which this application uses, gives access to the Twitter decahose, which is a random sample of 10% of all tweets. The Free plan allows for 5 million tweets to be returned to a single Bluemix account. The Entry plan offers access to all tweets, and allows users to create specialized tracks to more easily filter data.

For more information about the query language, go to the Bluemix Twitter documentation.

View your data using dashDB

Now that the data is in the app, you can also view it in dashDB, a relational SQL database. If you want to take a look at it, navigate to the tile for the service in the Bluemix Dashboard and click LAUNCH in the top right.

You can look at the Twitter data by clicking Tables in the left navigation bar, and you can run your own SQL against the data by going to Run SQL.

Analyze text with Personality Insights

The Personality Insights service analyzes text to determine personality traits of the author. It requires a minimum of 100 words, of which at least 70 must match words in the lexicon. To compute statistically significant estimates, however, a minimum of 3500 words are required, preferably 6000 or more.

For more information, go to the Personality Insights documentation.