Plotly Dash and OmniSciDB for real-time data visualization

I used Dash pretty extensively in 2018 (so bear in mind things have changed some since then).

It's a great tool but has some downsides compared to using d3, a chart library or even PlotlyJS directly.

* There is no good path to invert Dash and Flask relationship so that you can embed Dashboards into the larger context of a Flask application. Which really limits it for single standalone dashboards.

* The JSON serialization and dash components are hard to extend or customize as an end-user (without making changes to the library directly). So you'll quickly run into a barrier when extending for anything custom.

So it's a great tool, you can build responsive dashboards (that work with great Python libs such as Pandas for the processing) very quickly, but IMO its not suitable for embedding dashboards into other applications (which TBF isn't their goal).

I've used Dash professionally for the past few years and I think it's mostly terrible. It's really easy to pick up and you can get an interactive dashboard going quickly but past that it's useless. The API is completely undocumented, all you get is a few examples which you can't generalize to your use-case because none of their functions' parameters are described anywhere, on top of that they introduce breaking changes seemingly every version. Some plots are only discoverable through obscure forum posts by random users. Unless you plan to reverse engineer the library don't bother for long or medium term projects. In the end I was so frustrated I dumped it and made a custom dashboarding library which serves Matplotlib figures from a http server (everyone says their documentation is terrible but I think it's alright: it's complete and it's mature).