from Hacker News

Python Data Visualisation

by domoritz on 8/31/22, 4:06 PM with 7 comments

• by srvmshr on 8/31/22, 7:23 PM

  Wondering what is the purpose of the post. It is a cut-&-dry introductory post for Altair using very standard library features that most data science people (using Altair, albeit a smaller group than matplotlib) know. Maybe its just me, but I was expecting some neat hack using some less known features. What am I missing?

• by wilsonfiifi on 9/1/22, 5:26 AM

  Any time I encounter a charting or data visualization library, the first thing I look for is how difficult is it to export the charts to PNG. In most cases you’ll find out that they require you to install node.js to accomplish that. It will most likely involve rendering the chart using jsdom and then triggering the “click to export” button to generate the PNG file. This seems like such a convoluted process and I can’t help but feel there must be a much simpler/cleaner way to go about it.

  So far, Pygal [0] seems to be one of the few that gets it right in my opinion. The only drawback is that the charts aren’t as “pretty”.

  [0] https://www.pygal.org/en/stable/index.html

• by pizza on 9/1/22, 1:27 AM

  What is the go-to tool for realtime, 60fps animated visualization in Python? Preferably something versatile that could handle either spectrograms or tree-/graph-structured data.

  I’ve seen suggestions of using matplotlib animations but it’s never been smooth or streaming, in my experience. Is there is a great C/C++ library for this, also? It might not be too challenging to write bindings if necessary.