topic Re: Linnear regression in scatter plot in ThingWorx Developers

Linnear regression in scatter plot

jensc — Mon, 04 Mar 2024 12:47:22 GMT

Hello Community,

I was wondering if there is any way of "showing" linnear regression in a chart line?

I can only find the "HideLines" property on the line chart (scatter plot type) widget.

Here is an example using the g2plot library for further explanation: https://g2plot.antv.antgroup.com/en/examples/scatter/scatter/#line

Regards,

Jens C

Re: Linnear regression in scatter plot

Rocko — Mon, 04 Mar 2024 13:07:27 GMT

Hi Jens, is the goal to show the points and the regression line in one chart?

Then you would have to use the Combo Chart, if your version already has it. The Line chart widget can only show splines.

In the combo chart, you would have to provide the data for the regression line as well as the chart won't compute this on its own.

Re: Linnear regression in scatter plot

jensc — Mon, 04 Mar 2024 13:13:19 GMT

Hello @Rocko,

Yes, essentially I think that is what we would like.
Like this:

I did have a look in the Combo Chart documentation you linked, and the screenshot there doesn't exactly look like what we'd like. But perhaps we can work with it.
However, my 9.3.8 platform does not seem to include this widget... Is it something additional we need to install?

Regards,

Jens C.

Re: Linnear regression in scatter plot

Rocko — Mon, 04 Mar 2024 13:16:46 GMT

I'm afraid it's a feature of TWX 9.4+ only...

Re: Linnear regression in scatter plot

jensc — Mon, 04 Mar 2024 13:18:04 GMT

Hello @Rocko,

Ah okay, then I'll have to try it out on our 9.5 servers.
I'll get to testing it and get back if it is what I was looking for!

Regards,

Jens C.

Re: Linnear regression in scatter plot

jensc — Tue, 05 Mar 2024 09:42:59 GMT

Hello @Rocko,

I managed to get ahold of a 9.5 platform to try the Combo Chart.

However, I haven't really been able to create a chart similar to this one:

I do realize that there is no "Scatter" chart type for the Combo Chart.

I tried it using the following service just as a quick test:

result = DataShapes["ScatterPlot"].CreateValues(); const data = [ { regression: -2.2114, X: 1, Y: 4.181 }, { regression: -1.7274, X: 2, Y: 4.665 }, { regression: -1.0964, X: 3, Y: 5.296 }, { regression: -1.0274, X: 4, Y: 5.365 }, { regression: -0.9444, X: 5, Y: 5.448 }, { regression: -0.6484, X: 6, Y: 5.744 }, { regression: -0.7394, X: 7, Y: 5.653 }, { regression: -0.5484, X: 8, Y: 5.844 }, { regression: -0.0304, X: 9, Y: 6.362 }, { regression: -0.0124, X: 10, Y: 6.38 }, { regression: -0.0814, X: 11, Y: 6.311 }, { regression: 0.0646, X: 12, Y: 6.457 }, { regression: 0.0866, X: 13, Y: 6.479 }, { regression: 0.1976, X: 14, Y: 6.59 }, { regression: 0.3476, X: 15, Y: 6.74 }, { regression: 0.1876, X: 16, Y: 6.58 }, { regression: 0.4596, X: 17, Y: 6.852 }, { regression: 0.1386, X: 18, Y: 6.531 }, { regression: 0.2896, X: 19, Y: 6.682 }, { regression: 0.6206, X: 20, Y: 7.013 }, { regression: 0.4276, X: 21, Y: 6.82 }, { regression: 0.2546, X: 22, Y: 6.647 }, { regression: 0.5586, X: 23, Y: 6.951 }, { regression: 0.7286, X: 24, Y: 7.121 }, { regression: 0.7506, X: 25, Y: 7.143 }, { regression: 0.5216, X: 26, Y: 6.914 }, { regression: 0.5486, X: 27, Y: 6.941 }, { regression: 0.8336, X: 28, Y: 7.226 }, { regression: 0.5056, X: 29, Y: 6.898 }, { regression: 0.9996, X: 30, Y: 7.392 }, { regression: 0.5456, X: 31, Y: 6.938 }, ]; data.forEach((element) => { result.AddRow(element); });

Do you perhaps have an example I could check?

Regards,

Jens C.

Re: Linnear regression in scatter plot

Rocko — Tue, 05 Mar 2024 11:06:17 GMT

Nothing ready right now but I can check after my vacation.

Re: Linnear regression in scatter plot

Rocko — Tue, 19 Mar 2024 17:01:49 GMT

Jens, I have to roll back. Combo Chart isn't what I thought it was I was misled by the picture in the help, but you can't a line chart and a scatter plot into the same chart with it, sorry.

Re: Linnear regression in scatter plot

jensc — Wed, 20 Mar 2024 07:24:43 GMT

Hello @Rocko,

No worries.

Hopefully it will be added to the chart in the future.

Thanks,

Jens C.