[DataViz] Bins for histograms

Hey there 👋🏼

A lot of our users are starting to use Grist for generating reports for cabinet members (yay Grist ! 🥳 )

However, a fairly standard dataviz capability that is currently missing when making a histogram widget is the ability to define bins for the data to be grouped into before aggregating.

As of now, a workaround is to define a new “group” formula column and then use this column when making the histogram but imo that is not satisfying for two reasons :

1. this requires users to “code” which is not easy for some of them + feels hacky)
2. the order in which the bins are shown in the histogram is not easy to handle (for example, the following bins are sorted alphabetically but are not in the correct order : "[10-500]", "[10,000-100,000]", "[500-10,000]")

WDYT ? 🤔

[anaisconce]

Ah, yes! There is currently a way to do it with “natural sort.” When sorting a text column, there’s a “…” menu next to the column in the creator panel that has an option for natural sort.

But even with that, I can see the issue your team is running into. Adding the comma separator in the number is producing the wrong result. I propose fixing how natural sort handles numbers with comma separators in text columns.

I made a quick video re: what I mean for clarity.

https://user-images.githubusercontent.com/69366135/199968304-4cba50dd-b09b-4881-a633-356632049320.mp4

@berhalak may have some extra guidance about this?

[anaisconce]

Interesting! Would it also be useful to see the bins as rows in the aggregated table from which chart is generated?

That’s why I started thinking of summary tables first. If I understood correctly with the “hacky” method, that’s what the formula was essentially doing, right? Using a formula to categorize records into bins, then then aggregate the data by bin? Being able to do it from either the chart or the summary table UI makes sense, though the two are linked.