The process of data exploration can be rather cumbersome if the heatmap contains many rows. To reduce the complexity and to emphasize underlying patterns in data, user can, during a clustering procedure, reduce the number of heatmap rows by cutting the dendrogram and merging rows. Information about merged rows must be present in the InCHlib input file. inchlib_clust.py with the “-c” parameter can be used to prepare compressed data.

For example, to merge data objects with IDs “123”, “119” and “106” into one node “node@246”, the following syntax is used:

In the JavaScript code if the count_column attribute is set to “true” an additional color-coded column showing the count of items represented by one row is displayed.

In the example below original 176 rows (data objects) were compressed to 20 rows. A default InCHlib behavior is that using row compression will lose the information on the original data points. In this example, we demonstrate how this behavior can be changed so that the zoom-in view shows the original (uncompressed) data points. Using InCHlib events/methods and server side function the original rows are decompressed when zoomed-in and compressed again when zoomed-out (with the constant upper limit of 20 rows).