how to use histogram(data, num_bins)?

I have a 8945x1 double array, and I want to view it as a histogram

first try: histogram(data), here's what I get:

second try: using histogram(data, 20):

what am I doing wrong?

data is attached as an excel file if anybody cares to investigate

thanks!

[Walter Roberson]

Isn't this a duplicate with http://www.mathworks.com/matlabcentral/answers/278881-how-to-use-histogram-data-num_bins

Answer 1

[Ahmet Cecen]

You don't appear to be doing anything wrong. Your data is the problem:

 1417.69
 521.83
 15.23
 8.28
 7.05

The plot shows exactly what it should: and infinitesimal 1 count at around 1400 and 520, and a gigantic peak at around 0-15 range, since all of your data falls in to that range. Try:

 histogram(data(5:end), 20);

and you will be able to see what I mean.

[59morgan]

thanks for your help. I understand what you mean, about the data spread being a problem, but shouldn't increasing the number of bins give me a decent view? even up to maybe 50 bins? the idea is not to exclude the outliers but to see them alongside the others

[59morgan]

ok, so, what's going on is, it looks like just one big bin because there are many bins out to the right but I can't see them because they have such small counts... so how would you suggest I visualize this distribution, given that leaving out any data defeats the purpose? thanks

[Ahmet Cecen]

You can:

- Take log10 of your data and visualize that using histogram. Change your x label appropriately.

- Overlay them manually: take hist of data(1:50), then data(5:end). Crop and overlay the two figures. You will sacrifice accuracy but the message will get through.

- Manually plot the histogram as a bar plot, after using histogram to grab the counts. You can now get rid of the empty space between the 1400 value and 10 values, although x axis is no longer a coordinate system, simply a label.