As we've already stated, every Watson Analytics project that intends to analyze social media content must begin with defining a project as step one. Watson Analytics projects contain information about the topics and themes that you want to investigate. The project also covers the visual and text insights that are generated from the topics and themes you are interested in.

Depending on which version of IBM Watson Analytics for Social Media you are using (or your subscription level), the number and type of social media documents (such as Facebook and Reddit pages) that are supported varies. In addition, for YouTube content, Watson Analytics for Social Media can access up to 15 months of historical data. For all other content, 24 months of historical data can be accessed.

To control the volume of documents that is retrieved, carefully consider the topic keywords, context keywords, and exclude keywords that you specify. Also, check what time frame, sources, and languages provide the best results for the analysis that you want to do. Social media datasets contribute to your data limit in IBM Watson Analytics.

The entire content of social media documents that match the criteria that you establish is retrieved and analyzed. Information about mentions of your keywords in the documents that are retrieved is contained in your dataset and is represented in the visualizations on the Analysis tab.

To get started, in Watson Analytics click the arrow at the far left on the app bar, then click on IBM Watson Analytics for Social Media.

Note: depending on your version of Watson Analytics, access to social media might be slightly different.

The following screenshot shows the IBM Watson Analytics for Social Media Welcome page: