Create distribution of panel data for each year and industry

[albert master]

Hi!

I like to create a scatter plot where I can show simultaneously for 10 industries the average firm_size for each year by having multiple courses (sperated by different coulours) in one chart.
Is that possible in stata? Until now I was only able to create a distribution dependent by one variable :

bysort year: egen av_firm_size = mean(firm_size)

I thought about an if-operator but if I repeat this command it obviously says that I gernate the variable before:

bysort year: egen av_firm_size = mean(firm_size) if SIC_1==1

Do you have any advice for me that is replicable displaybale in the graph afterwards?
Many thanks!

[Andrew Musau]

10 may be too many to go back and forth from the graph to the legend. Here is some technique:

webuse grunfeld, clear
keep if company<5
gen yr= real(ustrregexra(string(year), "19(\d+)", "$1"))
collapse mvalue, by(company yr)
separate mvalue, by(company) veryshortlabel
set scheme s1color
tw dot mvalue? yr, vert xtitle("") xlab(35/54) msy(Oh + Th Sh)

Res.: