This small script takes input in tab-separated form (e.g. via copy & paste directly from Excel) and calculates cosine similarities between entities defined as lists of quantified variables (i.e. feature vectors). The output is a network file in GDF format where the edge weight is the similarity coefficient (multiplied by 1000 and rounded to an integer) between two entities.

Inputs need to have the entity names in the first row and the variable names in the first column. The easiest way to get a hang of this is to download this example file and simply paste the data in the textarea.

Source code available here.