Histogram
Six Sigma – iSixSigma › Forums › Old Forums › General › Histogram
 This topic has 7 replies, 6 voices, and was last updated 15 years, 4 months ago by Bonnie.

AuthorPosts

January 15, 2007 at 8:37 pm #45824
Hi,
Could some body explain how to calculate the bin size (class interval) for constructing a histogram.
Thanks in advance for your kind help.
Warm Regards,
Bon0January 15, 2007 at 8:57 pm #150593
qualitycoloradoParticipant@qualitycolorado Include @qualitycolorado in your post and this person will
be notified via email.Bonnie, Hello!Thanks for your posting. This thread from a few years ago covers this material, I think — see what you think:https://www.isixsigma.com/library/forum/c031022_number_bins_histogram.aspBest regards,QualityColorado
0January 15, 2007 at 9:08 pm #150595I usually take the range of the data and divide by 10 (cells). The I take whatever the interval came up to be and round it off to something that will give me reasonable cell intervals (e.g., whole numbers in an ideal world). The actual number of cells then is whatever that rounded number turns out to create.
0January 15, 2007 at 9:09 pm #150596I usually take the range of the data and divide by 10 (cells). The I take whatever the interval came up to be and round it off to something that will give me reasonable cell intervals (e.g., whole numbers in an ideal world). The actual number of cells then is whatever that rounded number turns out to create.
0January 15, 2007 at 10:16 pm #150602It is refreshing to see you are plotting manually. You will learn more this way rather than blindly using software like most people. Software can very much lead people astray, such as the idiotic normal distributions I see plotted over so many histograms.
Data Points Bins
50100 610
100250 712
>250 1020
0January 15, 2007 at 10:25 pm #150604Manually plotting histograms with an arbitrarily picked number of bins can very much lead people astray. You can create things that look like probelms with the data when no problem exists and you can visully hide things. Much worse than a little curve that is meant as a reference not a test.
Do you think that just because the guy wants to know how to do bins, he is plotting manually? He is trying to use Excel. Go create a histogram with Excel and see how the labels do not correspond to the data. Pretty idiotic.to use such functionality when so much better help is out there.
By the way the advice on bins is =sqrt(n) and adjust slightly to make logical bins.0January 18, 2007 at 12:59 pm #150729There is a very general rule called Sturgis’ or Sturges’ Rule with the following formula:
(Maximum value minus minimum value) divided by (1+3.322*log(n)) where n is the count of the number of observations. If you round this number it represents a fair approximation of bin size. This formula attempts to give as close to a normal distribution as the raw data will allow. You can then use the frequency function in Excel to create bins and categorize the data by bin.
Hope this helps.
Bill0January 18, 2007 at 8:16 pm #150759Thanks Bill,
Something like this is what I was looking for. This really helped.
Cheers,
Bon0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.