TRAC: Clusters Analysis

Pulsar’s network visualisation and segmentation capabilities have long been core to what sets our platform apart. With Clusters, you can transform any dataset or conversation into a dynamic network of meaningful groups, organised by their shared characteristics and behaviours. These clusters can represent authors, images, entities, topics, or keywords, giving you a powerful way to understand how ideas, people, and content connect.

Our clustering technology reveals patterns you can’t see in linear dashboards. By analysing the co-occurrence of data points, whether that’s keywords frequently used together, similar image concepts, related entities, or audiences with shared bio attributes, Pulsar automatically groups them into distinct clusters. This lets you:

Each node (circle) in the Clusters visualisation represents a specific data point, depending on the view selected:

The size of each node reflects its importance, determined by both how frequently it appears and how many connections it has. Nodes are automatically coloured into up to 10 distinct clusters, each named after the terms or entities that are most central within it.

Behind the scenes, Pulsar uses the PageRank algorithm to determine the relevance and influence of each node within the network. The final visualisation is arranged using a force-directed layout, meaning connected nodes are drawn toward each other, making patterns immediately clear at a glance

Clusters can be found within Audience Insights and Content Insights, under the following sections:

and each cluster is telling a unique story about what people are discussing, the type of content and media being shared, and the groups of people involved in that conversation. 

Below is an overview of how Clusters work across different data types.

You will find the Top Bio Clusters in the Audience Insights > Demographics section of your TRAC search. You can use this graph to identify the main groups of people involved in the conversation, based on how they describe themselves in their bios.

When viewing Hashtags in a Clusters Network chart, you can quickly identify groups of hashtags that frequently appear or are discussed together within the same conversation. This allows you to understand how different hashtags relate to one another and which themes or narratives they collectively represent. Pulsar automatically groups related hashtags into distinct clusters and applies a clustering algorithm to determine the relevance and influence of each hashtag within its group. This helps surface the dominant themes present in your dataset. Each cluster is then summarised to provide a clear, at-a-glance explanation of what that cluster represents.

When viewing Keywords in a Clusters chart, you can easily identify groups of terms that frequently appear or are discussed together within the same conversation. This reveals how specific keywords relate to one another and highlights the concepts shaping your dataset. Pulsar automatically groups related keywords into distinct clusters and uses a clustering algorithm to determine the relevance and importance of each term within its group, helping you surface the dominant themes in the data. Each keyword cluster is then summarised to give you a clear, at-a-glance understanding of what that cluster represents.

When viewing Entities in a Clusters network chart, you can quickly identify groups of people or organisations that are frequently associated or discussed together within the same conversation. This helps you understand how a brand or client is connected to specific individuals, institutions, or competitors. Pulsar automatically groups related entities into distinct clusters and uses a clustering algorithm to determine the relevance and importance of each person or organisation within that group, allowing you to uncover the most influential players in the dataset. Each entity cluster is then summarised to provide a clear, at-a-glance explanation of what that cluster represents.

Lastly, when viewing Topics in a Clusters network chart, you can identify groups of topics that frequently appear or are discussed together within the same conversation. This helps you understand how different topics relate to one another and which themes are shaping the narrative. Pulsar automatically groups related topics into distinct clusters and applies a clustering algorithm to determine the relevance and importance of each topic within its group, revealing the dominant themes in your dataset. Each topic cluster is then summarised to provide a clear, at-a-glance explanation of what that cluster represents.

In the Content section, we surface the top image captions, segmented and clustered into a network. The chart displays related captions grouped together into distinct clusters. By using this chart, you can distinguish different images that are associated with your brand, campaign, or a specific subject. Our clustering algorithm analyses each image cluster and determines the relevance and importance of each image within it, helping you to uncover the dominant and most important visual themes present in a given dataset.