This was originally Marco's idea. The idea is that there should be an "auto" setting, for how the histogram determines the number of bars to show. As the amount of data increases, so could the number of bars, so that more data leads to a higher "resolution" histogram.
Another method to calculate the optimal bin width / bin count of the histogram is Shimazaki and Shinamoto's method:
Histogram Bin-width Optimization
The page has code samples for Matlab, R, Python, Excel, and explains the method in detail.
This approach may appear more complex than various rules-of-thumb formulas, but it's easy to implement, and it produces great results even with multimodal distributions. Unlike Freedman-Diaconis rule, which in my experience is hit or miss, this method is insensitive to the scaling of data.
In a simulation I'm working on right now, there is a parameter with a peculiar multimodal distribution. This is how it looks like on histograms plotted according to different rules for the bin count. Please note how Shimazaki and Shinamoto's method reduces the bin width just enough to see the multimodal structure of the data distribution:
It would be nice if FlexSim histograms supported this method.
14 People are following this idea.
FlexSim can help you understand and improve any system or process. Transform your existing data into accurate predictions.
FlexSim is a fully 3D simulation software environment. FlexSim can be used to simulate any process in any industry.
FlexSim®, FlexSim Healthcare™, Problem Solved.®, the FlexSim logo, the FlexSim X-mark, and the FlexSim Healthcare logo with stylized Caduceus mark are trademarks of FlexSim Software Products, Inc. All rights reserved.
