A: ChartPixel handles multiple sheets, even if there are multiple tables spatially separated on a single sheet, but it processes each table independently. For example, if you have three tabs, ChartPixel will try to create charts from all three tabs having them processed and analyzed one by one, and then generate and display the most relevant charts from the combined pool of insights.

Now, you might ask: Why not try to combine or “join” these tables by some common key, if it exists? That’s a great question. The short answer is that no AI system, without deep context about how these tables were created in the first place and how they “conceptually” relate to one another, can reliably perform such joins. Should it be a left join, right join, inner join, or outer join? What about handling non-unique keys, inconsistent date formats, or overlapping time periods? These are critical decisions that can significantly affect the outcome and interpretation of your data.

When I worked as a data scientist, these were the kinds of choices that required careful thought as they had enormous gravity on the outcome and interpretation of your analysis. If this becomes automated, then too many things can currently go wrong (or we haven't yet figured out how to resolve this gracefully). That’s why we currently recommend that users join the tables themselves before uploading the data into ChartPixel. That said, we do plan to make this easier in the future—perhaps by detecting possible joins after upload and then giving users the option to confirm or adjust the join before proceeding. Thanks for the question!