When should I use a Python library like Plotly versus a standard BI tool?

All answers to this question.

Python libraries like Plotly are superior when you need highly bespoke visualizations that BI tools simply can't do, such as complex 3D surface plots, network graphs, or specialized scientific charts. Also, if your visualization is part of a machine learning web app (like a Streamlit or Dash app), staying in the Python ecosystem is much more efficient. However, for standard business reporting, BI tools are better because they handle user permissions, automated data refreshes, and mobile responsiveness much more easily than a custom-coded Python dashboard would for a non-technical audience.

   Answered 2024-11-16 by Pamela White

Since you mentioned Seaborn, do you find it frustrating that it isn't interactive by default? Does that lack of interactivity make it harder to sell your findings to stakeholders compared to Plotly's hover effects?

   Answered 2024-11-18 by Mark Thompson

• Mark, that’s exactly why I usually use Seaborn for static reports or papers and Plotly for anything that a user needs to "touch." Stakeholders love being able to hover over a data point to see the exact value. It builds a level of trust and engagement that a static image just can't match. If I'm presenting to a board, I almost always go with an interactive Plotly chart to keep them engaged during the presentation.

     Commented 2024-11-20 by Kenneth Hill

I find that Python is great for the "exploratory" phase of data science, while Power BI is much better for the "explanatory" phase where you're sharing with others.

   Answered 2024-11-21 by Michelle Young

• I agree with Michelle. Use the right tool for the right stage of the project. Coding for yourself to find patterns, and BI tools for others to see the results.

     Commented 2024-11-23 by Christopher Evans