Hacker News new | ask | show | jobs
Show HN: ipyvizzu – open-source animated charts in Jupyter Notebooks (github.com)
70 points by vidipeti 1536 days ago
7 comments

I work in the small team that created Vizzu (https://news.ycombinator.com/item?id=28895897), and now we've integrated our tool into Jupyter Notebooks to help data scientists and analysts present the results of their work easier.

ipyvizzu uses our open-source Javascript/C++ library, utilizing its generic dataviz engine that generates many types of charts and seamlessly animates between them. It is designed for building animated data stories as it enables showing different perspectives of the data that the viewers can easily follow.

Next to creating a Python API, we added extra features for this integration, like using data from a Pandas dataframe and auto-scrolling to keep the chart in position while executing multiple cells.

We would love to know what you think about it and how we should improve ipyvizzu.

Repo: https://github.com/vizzuhq/ipyvizzu Examples - click on any of them to see the code: https://vizzuhq.github.io/ipyvizzu/examples/examples.html#an....

Does ipyvizzu work well with Jupyter's slideshow functionality? (For those that don't know what I'm talking about https://medium.com/@mjspeck/presenting-code-using-jupyter-no... )
Not yet, unfortunately, I've opened an issue in our tracker for slideshow support: https://github.com/vizzuhq/ipyvizzu/issues/102
Great work. As someone who works on data products I am starting to come to the opinion that animated charts will become mainstream. If you haven't already check out Hans Rosling's work.
Thanks! This is our bet too. :) We built our tech from the ground up with this in mind. And we're huge fans of Professor Rosling's work, of course.

Our animated charts have proven to be quite popular for storytelling: https://www.reddit.com/user/VizzuHQ/?sort=top

But we have big plans to build self-service tools on the same core technology. We're currently raising our next round, btw.

The morphing of the charts looks great. I don't know how that works, but currently I am using eCharts to achieve similar effects.
The animation is brilliant! Good job!
Thanks :)
Looks great! Thank you!
Thanks a lot :)
Awkward name
The `ipy` prefix in the `ipyvizzu` name has a story:

We tried to implement a generic python integration first. It seemed easy with cython. But a question arised: does it work with interactive widget in jupyter notebook?

The `ipywidget` showed as a solution, but the custom widget was too complex for our usecase. So we looked for examples, and we found `plotly`. `plotly` has a nice trick: in jupyter notebook they call `IPython.display.display_html` with raw option. This way a html `script` node can be displayed in jupyter notebook.

`Vizzu` has a nice javascript API so we went in this way. `ipyvizzu` basically is a javascript code generator wich works only in jupyter notebook. That's why we put the `ipy` prefix for the package name.

The generic python integration is still a good idea, and we can use the `pyvizzu` or `vizzu` package name for it in the future.

Why?