Gaussfit (gaussfit)
The gaussfit
function fits a gaussian curve to a histogram constructed from a random sample drawn from a numeric field. The gaussfit function can be used to visualize how well the values in a numeric field fit a normal distribution. The gaussfit
function takes two parameters:

The numeric field from which to draw the histogram

The sample size
Sample syntax
select gaussfit(filesize_d, 50000) as fit,
hist_bin,
hist_count
from logs
Result set
The gaussfit result set contains one record for each bin in the histogram drawn from the random sample. The gaussfit
function returns the value of the fitted gaussian curve. The hist_bin
and hist_count
fields are also available in the gaussfit result set. The hist_bin
field contains the histogram bin number and the hist_count
field contains the count of samples in each bin.
Below is a sample result set in Apache Zeppelin:
Visualization
The gaussfit result set can be visualized by plotting the hist_bin
column on the xaxis and the fit
and the hist_count
columns on the yaxis. The visualization belows shows the gaussfit result visualized in an Apache Zeppelin line chart: