Doug's MATLAB Video Tutorials

Flow chart shows which visualization to use 4

Posted by Doug Hull,

Nathan over at FlowingData.com had a post that explains what visualization to use in different situations. Ultimately he linked to this chart from Andrew Abela of The Extreme Presentation(tm) Method

Click for original version
The version displayed below was modified with permission of Andrew.

MATLAB flowchart
Looking at this, I wondered if I could make all of those visualizations in MATLAB. The answer was yes, with a little help from MATLAB Central.

Download this .zip and run main.m to see MATLAB create all those charts with simple data.

vWhist
Spider
WaterfallChart
Histo2D

are included in the zip for convenience.

Many of these visualizations had dedicated functions to do that exact plot, some required some data massaging (like the “100% stacked bar plots” that normalized each stacked bar series to have the same sum) and some required the addition of some lines (like the connected bar charts).

This was a fun exercise in reproducing all these plots in MATLAB. Do you have some favorite plots that are not shown here? Do you want to make some dedicated functions to create those various bar charts that needed the most fiddling with?

4 CommentsOldest to Newest

I will post this group of files next week. I am hoping someone will make the bar chart mods and I would like to include them in the bundle.

-Doug

Hi.
This was a very inspiring post. It especially opened my eyes to polar plots for cyclic data which I work with regularly.

Unfortunately, it also opened my eyes to how hopelessly behing the matlab polar plots are to their rectangular cousins. The file-exchange provides some hot-fixes, but if you open up the results of any of them in the interactive plot editor you quickly realize the limitations.

–DA

There is a great variety, but what about combining grouped and stacked bar graphs? I haven’t come across that yet.. Thanks.

Add A Comment

What is 9 + 10?

Preview: hide

These postings are the author's and don't necessarily represent the opinions of MathWorks.