I came across your paper "Graph Layouts by t-SNE” from 2017 and found it most interesting! I was also happy to see that you had shared your code on GitHub. I’ve played around with the code a little bit and have a couple of questions that I hope you can answer:

• How did you achieve the edge bundling discussed in Sect. 4.7 and shown in some of the figures? Some post-processing of the saved graphs?
• I’m not sure how one could construct a PivotMDS layout for a new graph.

Göran Falkman

A self created file at the end of the run containing all node positions would save you losing all the data each time you close python. This can possibly be triggered using a --save flag?

Hi, I'm recently doing some experiments on your algorithm. (Thanks for your sharing!)
And I have a question about how to choose a good perplexity parameter. when I use your algorithm on other dataset, sometimes it will always say "The perplexity is probably too low". As is suggested, it seems it only has one proper value. I don't know what's wrong with the X matrix.

Hello.
I am doing my university homework, part of which is to implement some graph layout algorithm on the plane. I found your very interesting article about tsNET algorithm and wanted to try to implement it in my work. But there is no license specified in your repository, so I have no possibility to use your code. Could you provide me with such an opportunity? Or just specify the license in the repository. I will be extremely grateful to you.