I think I just found my new favorite search engine, Carrot2 (here). Centered around the visualization of search results using hierarchical pie chart clusters (more on that in a second), it not only displays results of your query but also the relationships between results. Let’s take a look:
So what is going on here is that:
[The] first ring represents root clusters and next levels represents their subclusters and so on. Size of each part on particular ring depends on number of documents in particular cluster.
In this way, they make interacting with large amounts of internet data more accessible.
Web clustering, the idea of defining the distance (or similarity) between two web objects through classification, allows users to see patterns and relationships between results. Clustering can be hierarchical, fuzzy or probabilistic, etc. but always implies grouping objects to increase their accessibility. The Carrot2 engine provides automatic organization of documents into thematic categories that you can explore textually and visually.
From their initial proof of concept, it has advanced to be a pretty powerful tool that also aggregates multiple search engine results. Did I mention it is open source? What I particularly like is that is allows for exploratory analysis of data quickly and easily using visual cues to symbolize hierarchy. I think this sort of pie chart could have significant applications for other types of data analysis using large datasets with hierarchically ordered data. Thoughts?
- Hierarchical edge bundles (seeingcomplexity.wordpress.com)
- How to Cluster a Generic Search Term Using Google’s Wonder Wheel and Image Swirl (searchenginejournal.com)