> I’ve worked with a lot of data visualization libraries over the years, both proprietary and open source, and I don’t know of any that can both display a million points this quickly, and support zooming, panning, and selection at interactive speed. (And now that I’ve I said this, I look forward to learning about some in the Hacker News discussion of this post )
Yeah, I should have clarified that I meant in 2D. ipyvolume is super impressive, and I didn't know you were working on bringing some of that webgl goodness to bqplot. Very cool!
Does ipyvolume have support some kind of points selection in 3D? Or is the idea that you'd perform selections using 2D projections in bqplot?
3.0.0 introduces a new Jupyter widget class: plotly.graph_objs.FigureWidget
For me, it doesn't. Every time I am tempted to use plotly I find it is either not working well with jupyter offline environments (a must for the data I am working with) or random bits an pieces from some official tutorial don't work.
Right now I have tried to follow the simple example that claims Jupyter. plotly.offline.init_notebook_mode and plotly.offline.iplot are no longer necessary -- which would be amazing. But when I try this, confirming I am indeed using plotly 3.0,
plotly.graph_objs
has no
FigureWidget
Maybe I am missing something obvious here. And I really would like to like this, but to me it's been a time sink more than a useful module.
Hi Jon,
Thanks for the great release!
Your post mentions support for JupyterLab, but getting Plotly working in JupyterLab is not working smoothly at the moment (see https://github.com/jupyterlab/jupyter-renderers/issues/132). Is this something you guys are looking into?
I haven't looked into this specifically. Using the FigureWidget approach in 3.0.0 doesn't actually require the current @jupyterlab/plotly-extension. It's all contained in the new plotlywidget extension.
Ah ya, I realized afterwards from more Googling that this is a global problem in the jupyterlab extension ecosystem when using node 10. Since it's already been worked on in other issues, I deleted my comment.
Excellent work!
From the article:
> I’ve worked with a lot of data visualization libraries over the years, both proprietary and open source, and I don’t know of any that can both display a million points this quickly, and support zooming, panning, and selection at interactive speed. (And now that I’ve I said this, I look forward to learning about some in the Hacker News discussion of this post )
FYI: https://github.com/maartenbreddels/ipyvolume does this in 3d already for quite a while, and based on this I've created a PR for bqplot (2d) https://github.com/bloomberg/bqplot/pull/661 That last one is not released though.
Thanks!
Yeah, I should have clarified that I meant in 2D. ipyvolume is super impressive, and I didn't know you were working on bringing some of that webgl goodness to bqplot. Very cool!
Does ipyvolume have support some kind of points selection in 3D? Or is the idea that you'd perform selections using 2D projections in bqplot?
np, yes it does do 3d selections (projected only for now), shown here https://twitter.com/maartenbreddels/status/96712877380723507... but not yet documented. Should be for the next version (0.5, to be released this year)
3.0.0 introduces a new Jupyter widget class: plotly.graph_objs.FigureWidget
For me, it doesn't. Every time I am tempted to use plotly I find it is either not working well with jupyter offline environments (a must for the data I am working with) or random bits an pieces from some official tutorial don't work.
Right now I have tried to follow the simple example that claims Jupyter. plotly.offline.init_notebook_mode and plotly.offline.iplot are no longer necessary -- which would be amazing. But when I try this, confirming I am indeed using plotly 3.0,
has no
Maybe I am missing something obvious here. And I really would like to like this, but to me it's been a time sink more than a useful module.
Edit: hn formatting
Do you have ipywidgets >= 7.2 installed (See install instruction at https://github.com/plotly/plotly.py)?
ipywidgets is an optional dependency and FigureWidget won't be available without it.
Thanks - that solved it! Works great now.
Hi HN,
This is Jon from the release announcement. AMA!
Hi Jon, Thanks for the great release! Your post mentions support for JupyterLab, but getting Plotly working in JupyterLab is not working smoothly at the moment (see https://github.com/jupyterlab/jupyter-renderers/issues/132). Is this something you guys are looking into?
I haven't looked into this specifically. Using the FigureWidget approach in 3.0.0 doesn't actually require the current @jupyterlab/plotly-extension. It's all contained in the new plotlywidget extension.
Oh, that's a great improvement in and of itself then. Thanks!
BTW, I got a GitHub notification of a comment that you were having trouble installing it. But I can't find the comment in GitHub itself.
Ah ya, I realized afterwards from more Googling that this is a global problem in the jupyterlab extension ecosystem when using node 10. Since it's already been worked on in other issues, I deleted my comment.
Hi Jon,
Just want to say I absolutely love plotly. It's made data science and visualization fun again. Thank you for all your hard work!
Sure thing! But full disclosure, I wasn't actually involved in any of the versions of plotly.py before 3.0.0 :-)
Hopefully 3.0.0 will be even more fun!
Plotly is awesome, but it can be really slow when loading from cold with a shiny app. Big fan of it with ggplot.
Is it possible to have the lasso tool return a list of selected data points?
Yes! You use the trace.on_selection method to register a callback function. Here's an example: https://github.com/jonmmease/plotly_ipywidget_notebooks/blob...
That's f'n awesome. Many thanks!