Gnuplot histogram

2
votes

I have created a histogram for some data that I have in a .dat file using the 

binwidth=...
bin(x,width)=width*floor(x/width)

plot 'file' using (bin($1,binwidth) + binwidth/2):(1.0) smooth freq with boxes

but because my file has over 1,000,000 entries that it reads from and puts into different bins, I don't know the exact frequency/density of each bin is. I would like to know the frequency or have it write the frequency to a file possibly, does anyone know if its possible? I was also wondering whether instead of frequency it was possible to plot percentage on the y-axis so example if my frequency is 50,000 and my total number is 1,000,000 then on the y-axis to have a percentage instead of the actual frequency?

Thank you

1
gnuplothistogram

1 Answers

7
votes

You can use the stats command to get the number of entries.

stats 'file'
binwidth=...
bin(x,width)=width*floor(x/width)

plot 'file' using (bin($1,binwidth) + binwidth/2):(1.0/STATS_records) smooth freq with boxes

To plot the percentage use

plot 'file' using (bin($1,binwidth) + binwidth/2):(100.0/STATS_records) smooth freq with boxes