[Python] Obtaining Dominant Frequency of a Time Series Data

I am working on accelerometer data and am trying to find the dominant frequency of my data. I use fft() function from fftpack in python. To find the dominant frequency I found the absolute value of the fft and found the frequency at which the magnitude is maximum. While doing this I get the dominant frequency to be around zero( 0-0.05 Hz). I've negated the DC content in the signal and found fft using

import numpy as np
import matplotlib.pyplot as plt
from scipy import fftpack

signal = signal - np.mean(signal)
freq = np.linspace(0.0, sampling_rate / 2.0, row // 2) # row:-no. of readings
signalFFT = fftpack.fft(signal)/row

plt.plot(freq, abs(signalFFT[:row//2]))
plt.show()


Even then I obtain the dominant freq to be around zero.

1. Is my method of finding the dominant freq and negating DC content correct?
2. What does a Zero Dominant Freq imply? Or in general what does a Dominant Freq = f Hz tell about the data?
3. What are the possible causes for Zero Dominant Frequency?
4. Is it possible for accelerometer reading to have a Zero Dominant Frequency when we know the data is being read from a moving object
5. What else can I do to get a Non-Zero Dominant Frequency?

Thanks a lot

Update 1:- Is there any rule that dominant frequencies of X, Y, Z axes of accelerometer should have a same or similar freq as Dominant Freq? If No could anyone please explain with an example.

Continue reading...