In this part of our Text Analytics tutorial we will learn how to extract key phrases from text messages with a little help from Microsoft Cognitive Services Text Analytics (AKA Azure Machine Learning). In the first part of the tutorial here we imported Clinton's and Trump's Facebook posts. It's time to learn how to extract the Key Phrases from those messages using Microsoft Cognitive Services Machine Learning, and with no help from IT, Developers or Data Scientists.

When we are done, we will be able to create stunning reports that will correlate key phrases with our metrics. In our example: We will be able to see which key phrases triggered the highest number of shares, and slice them by positive, neutral or negative sentiment, as shown in this report:

In my last community blog here we learn how to apply sentiment analysis in Power BI with Microsoft Cognitive Services API (aka Azure Machine Learning) and with no help from IT, developers or data scientists. In the follow-up articles we drilled down into different aspects of the Sentiment Analysis implementation in Power BI, and shared the technique in several community events. Following the great interest in these scenarios, in today's blog post we will use similar technique to extract key phrases from text messages. To make it interesting, we will pull Facebook messages from the public pages of Clinton and Trump, extract their key phrases, and correlate it with engagement levels and sentiment in cool visuals (As seen here).

You can really change up the look and feel of Power BI reports by utilising this simple technique or pattern. By adding a manual disconnected table and using the SWITCH function your report can take a different shape almost immediately.