Deploying Data Analysis Products

Hi, everyone.

I’m currently facing the following challenge on my job (a legal-tech-startup): our Data Analysis Team (aka me) can make the analysis, but we have problems deliverying the “product”.

We’re currently using Metabase, which operates with SQL unter the hood. It’s a great tool. We can write queries and it updates the results everytime it synchronises with our database. However, since SQL is a Query language, it does not deliver the flexibility that it’s possible to achieve with python.

For this reason, we’re looking for a tool/library/anything that allows us to embed JavaScript Charts in it. The library I use to plot visualisations (altair) reads my python code and converts it to JavaScript. This allows for some cool interactivity, but the only way I came up with to share the results with stakeholders is to share a html file of the chart with them, so they can open it in their browser.

Does any one has any recommendation on ways to deploy analysis made with python ?

Thanks in advance!

1 Like

Hi @celioxf

You asked me a question on another subject, but there I didn’t quite know the essence of your problem.

A fairly common tool is Apache Superset.

You can also convert your Notebooks to readable formats https://nbconvert.readthedocs.io/en/latest/latex_citations.html.

According to the Plotly documentation, you can pass JSON data to it - https://plotly.com/chart-studio-help/json-chart-schema/. But I’m not sure that building charts via Altair to pass to Plotly will be easier than passing charts directly from Pandas.

1 Like

The altair part is just because I’m already acquainted with this library, so it would be nice to be able to do most of the complex stuff in this library and simply pass the chart structure to plotly in a very early stage before I learn how to do it all with plotly.

I looked into Apache Superset and it looks interesting. If I understood it correctly, it runs with python and pandas under the hood (or at least you can use some pandas tools after extracting the data from the database.

Thanks for the recommendation!