Feature Reguest: ignore outliers

[yet-another-account]

With scatterplot, if there's a single outlier point, it can make the graph unusable. It would be great if there was a way to exclude datapoints in the top & bottom 5 percentile (configurable, of course) from scaling.

[etpinard]

HI @eukaryote31 thanks for writing in!

I'm assuming here that you're looking for a new scatter attribute. Your request feels similar to #299 where users have asked us to add a "jitter" attribute to scatter traces. In brief, some users would like scatter to include more "stats"-like draw settings (like box/violin points offer).

Personally, I'm not 100% convinced these settings belong in scatter. Deviating from the provided x/y coordinates in a scatter traces feels odd to me (at present).

Now, perhaps we could achieve this by extending our filter transform. You can already "filter out" points for example -> https://codepen.io/pen?editors=1010 - we could also way specify some aggregated field of the target data instead of a simple scalar. For example, maybe we could add support for:

Plotly.newPlot('graph', [{
  y: [1, 2, 3, 1e4],
  transforms: [{
    type: 'filter',
    target: 'y',
    targetstats: 'std'
    operation: '<',
    multiplier: 2
  }]
}])

which would filter out value past 2 standard deviation in y sample.

[gvwilson]

Hi - we are currently trying to tidy up Plotly's public repositories to help us focus our efforts on things that will help users most. Since this issue has been sitting for several years, I'm going to close it. If it's still a concern, we'd be grateful if you could open a new issue (with a short reproducible example if appropriate) so that we can add it to our backlog. Thanks for your help - @gvwilson