It was originally given as a 'job-test' exercise when I was applying for a job.
Also it was good excuse for me personally to learn more about:
- Python
- Numpy
- Pandas
- Anaconda
- Math: vector math, probability, DSP
Also I kind of like generating pretty charts.
It's time domain search of a specific signal and counting such signals. It serves no greater good, it was just an personal summer exercise. It's also inefficient & unoptimized.
Only if you want to know if I can write Python. Otherwise, no.
Finding sample in clean version of signal
Finding sample in signal with applied 25% of gaussian noise
Finding sample in synthetic version of signal (made out of samples)
Finding sample in synthetic version of signal, but with 33% of Gaussian noise applied
Processing steps on clean signal. First chart ORIG is sample. Second chart SIG is original signal. Other charts are processing steps.