Pandas Visualization for Data Scientists with D-Tale

[aschonfe]

My company, Man Numeric, has really been pushing for open-source projects and we were able to open source this client for visualizing pandas data structures. The motivation for building such a tool was born out of a conversion from SAS to Python and we needed a replacement for SAS's Insight function. It integrates really nicely into a jupyter notebook or python terminal.

Under the hood its a small Flask service with a React front-end (particularly react-virtualized & chart.js) which can be instantiated by passing it a Pandas data structure. It can handle datasets of any size (as long as the data fits in your RAM) by only loading & drawing what data is viewable in the dimensions of your browser. I have a demo running on PythonAnywhere:

http://andrewschonfeld.pythonanywhere.com/dtale/iframe/2

Let me know what you think?

I also have another heavier piece of data available for playing with:

http://andrewschonfeld.pythonanywhere.com/dtale/iframe/1

GitHub: https://github.com/man-group/dtale

Thanks

[ryan8020]

Got an error on the first example:

Traceback (most recent call last): File "/home/AndrewSchonfeld/.virtualenvs/dtale/lib/python3.6/site-packages/dtale/views.py", line 710, in get_data data = DATA[data_id] KeyError: '2Let'

[aschonfe]

Looks like your link copied a little but if the word on the following line of my post. It should be: http://andrewschonfeld.pythonanywhere.com/dtale/iframe/2

[ryan8020]

Thanks!