Example use cases for Amazon QuickSight - Discovering Hot Topics using Machine Learning

Example use cases for Amazon QuickSight

Example 1: View the sentiments, phrases, and tweets for a specific topic.

Select Twitter as the source, and then select a specific topic (for example, 004) in the table, to view the overall sentiment donut chart for that specific topic sentiment, the phrases in the most dominant topic word cloud, the heat map of tweets associated with the dominant topics, and a tabular view of related tweets.

Figure 7: Visuals in the second and third row render information for the selected Topic ‘004’

You can also use the Controls dropdown to expand the Topic and Sentiment options. These dropdowns have multi-select activated, thereby allowing you to select more than one topic or sentiment. Select topics that you would like to filter on to view the sentiments, phrases, and tweets associated with that selection.

Figure 8: Controls to multi-select topics or sentiments to filter visuals on the page

Example 2: View the and tweets with negative sentiments.

With Twitter as the source, from the Topic Analysis tab, select Negative. This refreshes the table of tweets in the third row visual.

Figure 9: Filter visuals based on the sentiment

Example 3: Select a phrase in the analysis and see the related tweets.

With Twitter as the source, from the Text Analysis tab, choose a phrase from the Phrases and their associated sentiment horizontal stacked bar chart. This filters the table of tweets that contain the selected phrase and the associated sentiment.

Figure 10: Filter tabular view of tweets with the selected phrase

Example 4: Filter tweets and sentiments for a selected entity.

With Twitter as the source, from the Text Analysis tab, select an entity from the Entities list and its associated sentiment from the horizontal stacked bar chart visual. This filters the table of tweets that contain the selected entity and the associated sentiment.

Figure 11: Filter tabular view of tweets based on the selected entity

Example 5: Filter for tweets and unsafe URLs containing associations with violent or explicit content.

With Twitter as the source, from the Image Analysis tab, select the unsafe moderation label filter the tabular view of tweets. The tabular view provides the date of the tweet, the tweet text, the tweet’s translation, and the image of the URL flagged to have unsafe content.

Figure 12: Filter tabular view of tweets in the second row based on moderation label selection