How to plot a swarm chart with no specific x-values?
显示 更早的评论
Hi everyone,
I have a huge dataset, around 1 million values in a column. The range of these values are limited. They, for example, fluctuate between -100 to +100 (I need to find these two values by min and max). The fluctuation is also very erratic (there might be some patterns though).
So, I was having an idea to use swarm chart and plot the distribution of this dataset on something like a swarm chart. Do you think it would be possible using swarm chart or any other chart type on MATLAB? (I know we can use histogram but something like what I described would be easier to understand for my audience)
I mean something like this:
5 个评论
Russel Burgess
2021-3-12
Wolfgang McCormack
2021-3-12
Russel Burgess
2021-3-14
I'm not sure if it's exactly what you're after, but you can just set the x values to all be 0, for example:
x = zeros(1e3,1);
y = randn(1e3,1);
swarmchart(x,y);
Produces:
Wolfgang McCormack
2021-3-14
Wolfgang McCormack
2021-3-15
采纳的回答
Russel Burgess
2021-3-15
Following on from the comments:
y = randn(1e3,1);
swarmchart(zeros(size(y)), y);
Will produce the desired chart, in R2020b.
For drawing a line on the chart at the widest point - unfortunately (as far as I know) swarmchart doesn't return or expose the plotted coordinates... publicly. That means you'll have to break into the structure and ignore some warnings.
figure(1);
clf;
hold('on');
y = randn(1e3,1);
s = swarmchart(zeros(size(y)), y);
raw_s = struct(s);
raw_data = raw_s.XYZJittered;
[xmax,maxidx] = max(abs(raw_data(:,1)));
ymax = raw_data(maxidx,2);
plot([-xmax xmax], [ymax ymax], 'r-', 'LineWidth', 3);
This may not be exactly what you want but it should be enough to get started: those last 5 lines of code expose the private scatter object fields, including the "XYZJittered" field which appears to contain the actual plotted data in 3 columns. I then just find the coordinates of the point with x value farthest from 0, and plot the line there. This yields something like:
This is a little crude - but once you've got the raw plotted data you could process it in other ways. Another option would be to calculate the kernel density directly to find the maximums (if you have the stats toolbox, or here but I haven't tried it).
更多回答（1 个）
This Github offering has worked pretty well for me:
5 个评论
Wolfgang McCormack
2021-3-12
Wolfgang McCormack
2021-3-12
Matt J
2021-3-12
I guess you figured it out (sinc eyou Accept-clicked the answer)?
Wolfgang McCormack
2021-3-14
类别
在 帮助中心 和 File Exchange 中查找有关 MATLAB 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
