How to generate a pareto plot with an ordered X-axis?

Hello,

I'm looking to organize my data into a plot similar to a pareto plot, but I want the x-axis to be in ascending order from zero like a normal scatter plot. Another way to describe what I am looking for is a histogram (where the x-axis is various buckets of values) but with the y-axis being the sum of the values in each bucket instead of occurences.

I attached some of my data, basically I want to be able to plot frequency on the x-axis, and the sum of column B on the Y axis. Column C is a running sum of Column B, and its basically this data that I want to see in bar form.

Is there a way I can modify the pareto or histogram functions to do what I want or do I need to do something from scratch?

Any insight would be appreciated.

Thanks,

Matt

Answer 1

[dpb]

xyz=readmatrix('sample_data1.csv');
whos xyz

  Name       Size            Bytes  Class     Attributes

  xyz       61x3              1464  double              

xyz(1:10,:)

ans = 10×3

184.2248 0.2222 0.2222 187.2484 0.0075 0.2296 192.4868 0.0000 0.2297 201.2158 0.0000 0.2297 225.6160 0.0000 0.2297 230.3885 0.0019 0.2316 238.8880 0.0000 0.2316 263.4815 0.0012 0.2328 265.9375 0.0040 0.2368 291.1873 0.0101 0.2469

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

numel(unique(xyz(:,1)))        % how many unique values of first column are there?

ans = 61

Well, they're all unique.

I'm not at all sure what, specifically, you envision the plot to be or how to interpret "plot frequency on the x-axis". What does "frequency" mean here?

One guess would be simply

bar(xyz(1:10:end,1),xyz(1:10:end,3))

●

that assumes equal number of rows/bin and the associated cumulative total at that bin location. One could even out the x-spacing by using

[~,edges,bin]=histcounts(xyz(:,1),nbins);

and use the last element of each bin edges to get the x-position associated with the bin to get the accumulated total for the bin.

I don't believe you could make the Pareto plot do anything like you want; it computes the distribution from the inputs regardless, you cannot change its internal interpretation of what the data represent.