With duplicated data, area mark goes a little nuts, line mark is fine

[pdebruic]

When plotting a time series with multiple instances of two (or more) data points at the same time the area plot doesn't draw a coherent shape.

This helped me find a bug in my own stuff. So I'm fine after removing my duplicated data.

This is an example
https://observablehq.com/d/bd4e9b56a1163065

[mbostock]

Can you make the notebook public so we can see?

If you’re stacking the data, you’ll need to make sure the x-values are aligned. Often this is done using the bin transform (or group transform) prior to stacking.

[pdebruic]

Oh sorry. Its public now. I was not trying to stack. My bug for my thing was I was mistakenly trying to plot the same value on the same date, multiple times. Based on what was drawn it made me wonder if the area plot should've behaved in a "sane" way like the line plot.

[mbostock]

It’s a consequence of the implicit stacking with areaY when you use the y channel. If you use the y2 channel instead (to set the topline of the area) then you can opt-out of the implicit stacking.

Plot.plot({
  marks: [
    Plot.areaY(toPlot, { x: "xVal", y2: "yVal", opacity: 0.3 }),
    Plot.line(toPlot, { x: "xVal", y: "yVal", tip: true })
  ]
})

[pdebruic]

Oh. Thanks! I didn't understand what was going on. Thats clear.

…

On Oct 9, 2024, at 6:41 PM, Mike Bostock ***@***.***> wrote: It’s a consequence of the implicit stacking with areaY when you use the y channel. If you use the y2 channel instead (to set the topline of the area) then you can opt-out of the implicit stacking. untitled.-.2024-10-09T184027.405.png (view on web) Plot.plot({ marks: [ Plot.areaY(toPlot, { x: "xVal", y2: "yVal", opacity: 0.3 }), Plot.line(toPlot, { x: "xVal", y: "yVal", tip: true }) ] }) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>