How to graph an epidemiological cluster?

[Francesco Bianchi]

Hi everybody, I'd need assistance please.
How can I graph on Stata an epidemiological cluster similar to the one attached?
Can somebody help me please? Thanks in advance

Francesco

[Mike Lacy]

I have two thoughts: First, I would not recommend Stata as a tool here. I don't think of it as especially useful as a generic graphical drawing package. I'd think that the graphics features of an ordinary word-processing package would be better and easier. I rarely do that sort of thing, but people who do can likely recommend software. Second, one can use the features of Stata's SEM builder as a graphics editor for some parts of drawing arrow diagrams, but that seems like a hard way to go.

[Leonardo Guizzetti]

This sort of graph can be built by layering elements of twoway graph types (-scatter- and -pcarrow-) but this particular arrangement looks excessively tedious and built for purpose.

I guess the y-axis is meaningless, but does the x-axis have an implied meaning of time? I guess I don't find this type of graph that convincing for conveying a cluster, as you call it. This data could be cast into this more traditional graph, with the bonus of meaningful axes, and judicious choices of colors/annotations can add the same contextual information.

[Andrew Lover]

Hi Francesco,

You might be able to reverse-engineer something using nwplots from nwcommands (SSC)....

https://nwcommands.wordpress.com/tag/graph/

But as Leonardo suggests, a color-coded epicurve with transparency might work just as well.