I am using a Pulse Counter to send pulses from a water meter to Ubidots, and convert the pulses to a flow rate.
I noticed that my pulse counter is somewhat sending off pulses in rapid succession, which contribute to these erratic and large spikes in the flow rate (which I am converting using a synthetic variable).
I have tried different attempts at smoothing and filtering out these spikes, as they are definitely not an actual representation of what is happening. I have not had much luck using various timeshift, rolling means, or where statements.
I’m hoping someone has some ideas using Synthetic Functions to smooth these out; ideally I simply do not include these spikes in my data, or substitute them for previous “normal” values or a rolling average.
I’d like to delve deeper into the issue you’ve raised. Specifically:
1 -You mentioned making several attempts to smooth and filter out data spikes. Could you provide more details on the methods and parameters you’ve tried?
I have tried different attempts at smoothing and filtering out these spikes, as they are definitely not an actual representation of what is happening. I have not had much luck using various timeshift, rolling means, or where statements
However, can you elaborate please?
2 -What are your criteria for determining accurate data? For instance, do you have a specific range you expect the data to fall within? Additionally, could you specify the acceptable spread or standard deviation for the data points?
