High-quality graphic representation

[Julian Ruf]

Hello my dears,

I need a few creative minds and ideas.
My research relates to the influence of income on the chance of being overweight.
1.
I would like to create a graphic for an academic paper that represents the relationship between obesity and income.
Basically I am working with the variables (overweight => binary) and (income quintile = categorical) in my regression. However, it would be conceivable to use the continuous form of the variables (Income and BMI) for graphical representation. The differences in the BMI distribution in the quintiles are small, so I don't think a boxplot is very meaningful.

2. I have a panel data set over 15 survey waves, and would like to present temporal differences in the variables.

Do you have any ideas on which graph(s) are suitable for a bachelor thesis?

I look forward to your answers

[Nick Cox]

Please try to run one or at most two threads at once, not several as you are currently doing. This one may or may not get answers beyond mine now, but in my view it's not well pitched as a Statalist question.

Statalist works best with a highly specific question that is pretty clear to experienced users given a context centred on code and data that you show us and explicit detail on what you want, e.g. to solve a problem that you explain.

This looks more like a fishing expedition. Depending on local arrangements, you may well have a supervisor, advisor, mentor, whatever, someone whose job it is to provide advice. In my local context, this kind of question would call for a supervisory meeting in which I as supervisor would ask several questions and I would want to look at the data and ask what you have done so far. That could easily be quite a long meeting and it is hard to replicate here. If some kind of one-to-one support is not allowed or not available where you are, it's not Statalist's function to act as a substitute.