PyData 2021 - Build polished, data-driven applications directly from your Pandas or XArray pipelines

PyData 2021 - BUILD POLISHED, DATA-DRIVEN APPLICATIONS DIRECTLY FROM YOUR PANDAS OR XARRAY PIPELINES

Philipp Rudiger | 10/29/2021 12:00 PM UTC - 10/29/2021 12:30 PM

What if building a data-science app could be as easy as replacing arguments in your data pipelines with widgets? Panel and hvPlot .interactive allow you to leverage the APIs you already know (Pandas, Xarray, Dask, cuDF, etc.) to build everything from one-off interactive applets all the way up to full-featured polished web applications.

Notebook, Video

image

@philippjfr

4 Likes

Congrats @philippjfr on the presentation. I think .interactive is something truly new, useful and powerful. I hope it will get a lot of users.

I even got so inspired I built a small app using hvplot .interactive.

Check out my showcase Hvplot .interactive - A truly amazing api for making interactive data exploration and data apps from Pandas DataFrames - Showcase - HoloViz Discourse

2 Likes

Can .interactive cleverly understand what parts of the pipeline to recalculate? Do it use some kind of caching? it just seems so fast and smooth @philippjfr .

1 Like

Yes, it caches each step of the pipeline and only recomputes the parts downstream from the parameter that changed. It’s different from existing HoloViews and Panel pipelines in that it evaluates eagerly. This means that it always caches the most up-to-date result based on the parameters/widgets that feed into the pipeline. This is required so that we can accurately mirror the API of the object that is being wrapped. For very expensive calculations it might therefore not be entirely suitable but in those scenarios you could switch to dask, which would ensure that the actual execution is still lazy.

2 Likes