Hi,
I would like to create a single geographical map of two variables to visualize how they vary across the study area/country. Basically, in my analysis i look at the impact of CDI (this is an index ranging from 0 to 1), on TE (this is also an index ranging from 0 to 1). What i want is a map of the study area showing CDI and TE distribution by geographical units (in my case districts) to visual the variation of CDI and TE across the study area. So far: I have the district-level administrative boundary GIS data and was able to generate map of CD and TE separately using spmap in stata, so far now i pasted these maps on top of each other.
Is there another better way around this? Thanks a lot.
Apologies if my answers aren't useful, since I'm not familiar with stata or your particular field of research, but your question is interesting.
One idea might be to plot some combination of CDI and TE, e.g., (CDI + TE), (CDI*TE), or (CDI-TE), or even something more exotic like (2*CDI - TE); one or more of those might show some interesting patterns, or support/contradict some particular theories.
Another idea might be to plot the CDI values using a colour index, and then print the TE value for each district as a number on top. Not ideal either, I know.
My natural instinct would be to plot one variable using a "shading" method, i.e., a palette of colours with a single colour for each range of values between contour levels. Then on top of that, plot the other variable using contour lines. This kind of thing is done on weather maps all the time (see, e.g., wxmaps.org), with say temperature shown as colour shading with pressure shown as line contours.
One problem with that approach is that your data are based on "political" regions rather than being on a regular grid, and you may lose some information by interpolating your data onto a regular grid, with contours cutting across your distinct regions. On the other hand, interpolating your data onto a regular grid and plotting the resulting pattern may prove quite revealing in its own way.
Another problem is that I don't know if stata can draw such plots anyway. At least you can get over that problem by using a different plotting package, e.g., GrADS (http://cola.gmu.edu/grads/), which is very powerful, easy enough to use, and free. The downside is that GrADS really expects the input data to be on a regular grid, and colouring in just your political districts will be difficult with GrADS, even if possible at all.
That's my 0.02c worth in any case. Good luck!
Dear Enda William O'Brien,
Thank you for your time and the detailed helpful answer. indeed, your idea of the "shading" method is appealing and something i was also thinking about, but i could not manage to do it with Stata. It is either that stata can't do it or probably i am the one that can't get the stata to do it for me. In my case i thought i am probably the problem and there is probably someone to help with that. But it seems like stata is not capable of doing something like that (at least i am not able to figure it out yet). I will try to use GrADS as u suggested there, and see if i can produce something like that. Adding up the two variables don't make sense, well at least i cann't make any sense of it. my data are based on geographical regions (i.e. i would like to show CDI and TE across the 389 districts), apologies if i did not communicate that well. Thank you for your insights, let me try the GrADS thing and see what happens.
Hayat
- Badges
- Science topic
- Similar topics
- Economics