[FEATURE]: Support hovering all overlapping points #8113

Activity

1. Lexachoc

   added

   featuresomething newsomething new

   on Oct 9, 2026
2. camdecoster commented on Oct 13, 2026

   

   camdecoster

   on Oct 13, 2026

   Contributor

   More actions

   Thanks for the write up! Could you provide an example where you think this type of feature would be useful? How does your request differ from the previously filed issues?
3. Lexachoc commented on Oct 14, 2026

   

   Lexachoc

   on Oct 14, 2026

   Last edited by Lexachoc

   ContributorAuthor

   More actions

   The main difference is that I am not necessarily proposing multiple native hover labels.
   What would be most useful to me is exposing all records that overlap at the hovered position, so applications can decide how to present them.

   For example:

   x: [1, 1, 1]
   y: [2, 2, 2]
   customdata: [
     ["measurement A"],
     ["measurement B"],
     ["measurement C"]
   ]

   In my use case, customdata may contain metadata describing how each measurement was produced, such as processing/manufacturing conditions, provenance, and measured properties.

   At the moment, hover picking returns only one of these points — the current hover winner. It would be useful to preserve that existing winner for backward compatibility, while also exposing the other points that overlap it.

   With access to those records, an application could decide how to present them — for example in a grouped custom tooltip, an expandable/pinned tooltip, or potentially a native plotly hover UI if such support were added later.

   One concrete use case is experimental materials data. Multiple measurements can have the same nominal composition but different measured property values because they were produced or measured under different conditions. Being able to access all records at that composition would let users compare those measurements instead of silently seeing only the hover winner.

   This also comes up in my recent focus on contourternary. In experimental materials datasets, several measurements can share the same composition while having different property values due to different processing or measurement conditions.
   This is of course a separate issue from how those values should be handled during interpolation. However, when the original experimental points are displayed on top of the interpolated contour, being able to inspect all measurements at the same composition would be particularly useful.

   P.S. The current workaround is with some data prep, for example by pre-grouping overlapping records in customdata or an external index.
4. KoolADE85 commented on Oct 14, 2026

   

   KoolADE85

   on Oct 14, 2026

   Contributor

   More actions

   Does a hoverdata event not contain all the data points already? I'm wondering if we could support custom implementations of hoverdata by just providing all the raw data.
   Or is this issue strictly about visually showing all hover points by default?
5. Lexachoc commented on Oct 14, 2026

   

   Lexachoc

   on Oct 14, 2026

   ContributorAuthor

   More actions

   Does a hoverdata event not contain all the data points already? I'm wondering if we could support custom implementations of hoverdata by just providing all the raw data.
   Or is this issue strictly about visually showing all hover points by default?

   I would ideally like both: exposing all overlapping points in hover data, and a native way to display them. Similar behavior can already be implemented with some data preparation, for example by grouping overlapping records in customdata, but exposing the raw overlapping points would make custom hover implementations easier.

   Codepen: https://codepen.io/Lexachoc/pen/KwWomLQ

   ---

   I also understand that the current hover label design was not really built around displaying overlapping records. If I were implementing this today, I would probably disable the native hover label with hoverinfo: "none", listen to the hover event, and render a custom tooltip myself.

   This example (https://codepen.io/Alex-Hsu-the-bashful/full/EayrEyQ) also gave me some inspiration for that approach. (example from: #7707)

   Codepen: https://codepen.io/Lexachoc/pen/yyMKXBE

   👍React with 👍1Reacted by Cameron DeCoster
6. camdecoster

   added

   P3backlogbacklog

   enhancement

   size: 5

   on Oct 15, 2026