FFT plot: plot arrays in JSON

[dsxoiniana]

I’m checking the first option with the time series bucket in Heatmap. Below is the result. Indeed, each line in the diagram corresponds to one column name for the magnitudes, but with proper renaming this could be mapped to frequencies. However the deviation is not visible that well as in the previous diagram which has the y axis in logarithmic scale. Is there a way to apply a similar scaling in the time series bucket mode?

By the way, I looked at the advanced variable options in the Grafana documentation and if I modify my query to something like this:

SELECT ${amplVar:doublequote} as yAxis, ${freqVar:doublequote} as xAxis FROM "meas_ext" WHERE ("deviceID" = 'CurSensor_MCU') AND $timeFilter

I’m now able to dynamically select and add plots on the same diagram.

[zuim]

I didn’t know about the advanced variable options, thats good to know!
Regarding the scale of the buckets I think they are always evenly spaced. You could save the data in a logarithmic scale, when collecting it to have columns already spaced correctly, although that seems like a kind of bad hack.

An alternative is using the plotly panel I menttioned above, which allows log scale for y and for the data color. I tried it using some different data so you can get an idea how that would look like:

It does not have the nice even spacing of the Heatmap panel, but with always the same frequencies for y it should look a lot more uniform.