Topics in Mathemagenic (2)

This post is an example of struggling with a lot of data and trying to find the patterns. My favourite technique in such cases is to visualise it, hoping that the patterns I would expect in the data are clearly visible.

Looking further at the relation between (manually assigned) topics in the content of weblog posts, it seems reasonable to assume topics have a broader / narrower relation. The figure below is an example of visualising this. Top-to-bottom are the topics and left-to-right are the topics in the same order. The colour scheme is from red (hot) to blue (cold). An example, knowledge mapping has two red broader topics, namely: knowledge representations and ontologies. Further down, personal knowledge management, blog research and PhD have a lot of red, suggesting they are linked to broader topics Lilia is interested in. So far so good, but I would have expected favourite topics to cluster at the top.

Next step is to figure out whether there is something wrong with the data, the algorithms, or the visualisation. Research is hard work :-).