Topic Analysis allows you to go beyond the volume of results or where conversations are taking place and drill further into what the general topics and tenors are related to results.

This provides you with the ability to find deeper context for the results you see and better understand how the results are being talked about within separate topics.

Topic Analysis is a complementary tool to be used with the Overview and Analytics tabs in Explore.

This article will cover:

The chart shows the topics that appear most frequently. It has three levels of topics and subtopics. Click on a topic to expand the subtopics.

Note: You can select the third level of the topic wheel for a topic drill-down. An AI-Powered Insight and post previews will appear.

Analytics included in the drill-down:

Context Filter:

Adding a topic as a filter will create a Context filter named after the topic and will filter your search results. Context filters are automatically applied when drilling into Explore metrics that don't have corresponding filters. They cannot be directly created or edited.

The size of a spoke on the topic wheel is relative to the size of the overall data set; regardless of size, it presents an opportunity for analysis.

This widget provides a total volume of all mentions from all sources in your selected date range. It also includes the delta trend, which tells you the percentage change in volume over the previous time frame.

Topics are automatically assigned to documents using natural language processing (NLP).

This widget features six more word types, including Hashtags, Products, People, Organizations, Locations, Emojis, and Keywords.

Words are color-coded by type:

Word size in the cloud is dependent on the frequency of results. The more occurrences of the word, the larger it appears in the cloud.

Hover over any item in the cloud to see the total number of occurrences.

Click on the legend to include/exclude word types from displaying in the cloud. When excluding items, the cloud will dynamically re-fill with items from the other still-selected types, by occurrences.

Everything selected:

Only Location and Emoji selected:

This widget shows positive and negative keywords from mentions that appear most frequently in results.

This widget shows the overall tonality of mentions in your results. Sentiment is derived from our natural language processing algorithms. The sentiment of a mention will either be rated Positive, Neutral, Negative, or Not Rated. A mention will be labeled as Not Rated if it does not have enough text to analyze sentiment.

To drill down by sentiment, click on a sentiment type, and it will apply the sentiment filter to the query.

To read more about sentiment and how to override sentiment, click here.

This widget shows the tonality of mentions over time. Sentiment is derived from our natural language processing algorithm. "Not Rated" mentions don't have enough text to analyze sentiment.

This widget shows the tone of mentions by source. Sentiment is derived from our natural language processing algorithm. Mentions labeled as "Not Rated" do not have enough text to analyze. This widget can help you explain why you focus on one source during a certain campaign or why you put more effort towards a certain source.

This widget shows the object and scene categories detected most frequently in images that are posted or attached to posts. Only News and Reddit are supported.

This widget shows the top hashtags associated with your query.

Hover over a top hashtag to see how many times it appears in your results. Click on a specific top hashtag to see all mentions for it. Explore will add a sub-keyword filter with the hashtag you selected to the query.

This widget shows how many mentions in your results express a specific emotion. This differs from sentiment analysis, as emotional comparison is based on specific keywords, phrases, and emojis. This includes negation (keywords that won't be matched).

E.g. "I love this" vs "This made me sad".

The emotional comparison widget currently supports keywords and phrases in English. However, it will return results for other languages that have a specific emoji in the mention.

Note: YouTube data is excluded from all of the above visualizations per their policies.

Each widget can be exported by selecting the 'Download" Icon, Download CSV in the upper right-hand corner of each individual widget.

This ‘in context’ view allows you to quickly view the mentions for any metric without altering your search results via boolean, filters, or refreshing. Filtered mentions are available within the Topic Analysis tab, as shown below:

Filtered mentions allow you to uncover the mentions driving metrics with a single click.

Within Topic Analysis, click on any* word or chart to open a new content stream to the right of the screen that automatically filters for only those results.

The Filtered Mentions content stream features search and boolean tips by showing what filters or boolean strings can be applied to your search to update your total search results to only those in the Filtered Mentions content stream.

The Filtered mentions content stream includes all the same features as the main content stream, including search, bulk actions, filtering, downloads, and more!

Learn more about other Explore Widgets:

💡 Tip

Need more help? Feel free to reach out to us via Live Chat or check out our Customer Community.

Find answers and get help from Meltwater Support and Community Experts.

​