For my project I need to evaluate large amounts of wlan signal strength measurements.
Measurement is in dBm which is a logarithmic scale for milli watt (so every 3dBm the milliwatts double) where 0dBm is 1 milliwatt and -100 dBm is 0.1 pW. Received Wlan Signal Strengths are usally from -30dBm (very good) to -95 dBm (very bad).
These 2 examples were one Router measureing another's received signal strength:
First is the measurment over time; the red line is a naiive linear regression line. The blue and yellow plot bands are the 60% and 80% confidence intervalls; green dotted line is mean
The second graph shows the frequency distribtuion and the red line is the CDF of a normal distribution with given mean and std dev
Distribution 1: (Samplesize: 500, Mean: -49.51, Median: -49, Mode: -49, σ: 1.77)
Distribution 2 (Samplesize: 500, Mean: -64.41, Median: -64, Mode: -64, σ: 3.02)
Practically everytime I get mean = median = mode (with +/-1) but a nice distribution curve that fits a normal distribution CDF only with high measurements.
My question is how would you classify this distribution? I heard log-normal but a CDF did not fit any of my distributions.