Segmentation of Data Set into Stationary and Non-Stationary pts

Dear everyone,

I have a massive Dataset from an Experiment with stationary and dynamic points in it. I would like to segment this Data set into Stationary and non-stationary points (or Dynamic points)
The Idea behind this is to create seperate files for the stationary and non-stationary points.

I have tried a few things, including the mean/std etc.
Could anyone of you, point me into the right direction or library that would help me achieve this?


In image above you can see a snap of the Data.

The red lines depic a change of the Blue line of 0.5. However this doesnt seem to capture all pts as seen in a zoomed image here (cannot upload more than one picture.) :

Thanks for all the help!

Have you got these data in a pandas DataFrame?

If so, has you considered using diff?
https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.diff.html?highlight=diff#pandas.Series.diff

Suppose your raw “blue” data are in a column named “blue” in the
DataFrame df:

 df["blue_delta"] = df["blue"].diff()

then the dynamic points are where df["blue_delta"] != 0. You can also
do a compare on a whole series, eg:

 df["blue_delta"] == 0

which would match the staionary data points. And since the above is a
Series of Boolean values (true/false) you can use it as an index to
select the “stationary” points and save them in their own column:

 df["blue_stationary"] = df["blue"][ df["blue_delta"] == 0 ]

All the above is untested, but hopefully can point you in a useful
direction.

Cheers,
Cameron Simpson cs@cskk.id.au