Averaging data steps

[ritch_by_nature]

Hello,

        I have this set of data that has 9 steps in it that I am trying to analyze but before each step there is a big jump. I am trying to get the average of each step and ignore the peak of the jumps. I dont want to do it manually because I have other data with diff step length.

Is there anyway I can go about it to get the average of the first step, ignore the jump then go to the second step and average it, and so on. Please refer to my pic for a better picture!!

[nyc_(is_out_of_here)]

There may be other more elegant ways.

I would take the value in the middle of each of the steps and throw out anything that is not within 10%.

[ogk.nz]

I would say go through the data with a peak detector, find the peaks and given that you have a pretty good idea of the with of the spikes, exclude all data within X of the peak, this may able to be done after going over the data a first time with a short running average to try and smooth out the roughness of the steps.