Hi Folks:   I'd like to describe the extent of inequality in health behavior in Korea using descriptive statistics.The variable is continuous and the data is attached to this post. I already mapped actual health behavior using proc gmap approach shown in the code below.    But, I also would like to map the standard deviation and inter-quartile difference of each observation (each district and 250 districts in total) of the variable using the same gmap approach. Since mapping IQR diff and STD maybe more meaningful than visualizing actual variable given the purpose is to show disparity in the spatial distribution of the variable.    Could you please help me compute inter-quartile difference and STD of the variable for each observation in the data?  I greatly appreciate any hints or suggestions.  Thanks in advance.    proc gmap data=have map=mapsgfk.south_korea; id id; choro health_behavior/ legend=legend1; legend1 position=(bottom center inside) across=1 mode=reserve label=(position=top j=c); format std_health_behavior comma5.; run; quit;     ... View more

Hi Folks: Is it possible to define title in Macro for proc gmap? I have too many maps and tedious to copy and paste title with each call. Defending the purpose of macro. Mock data attached.  Thanks in advance.    %macro maps(var=,label=); proc gmap data=mydata=mapsgfk.south_korea all; id id; choro LOCAL_ECONOMY / legend=legend1; run; quit; %mend maps; TITLE height=11pt FONT='Arial' 'Economy'; %maps(var=local economy,label=Economy); ... View more

Hi Folks:   I'm trying to produce a map as compact as possible with minimum blank space around and legend tucked in the way it's closest can be to the body of a map as shown in the Desired Output. Legend could be placed anywhere as long as it's space efficient. My macro produce 11*3=33 pieces of maps with blank space around and legend further away as shown in Current Outout. Using proc panel jointly with my macro is not possible. But I still have to insert these maps in the table cells in Word.doc to create a 11*3 paneled view. I attempted to copy and paste the Current Output shown in the image below in Microsoft Paint and manually produced Desired output. However, doing it 33 times would be tedious plus pixel of images reduce.  Your advise to produce Desired Output map is greatly appreciated. Please see the mock data attached to this post and the proc gmap below int he snippet.    Thanks in advance.      proc gmap data=mydata map=mapsgfk.south_korea all; id ID; choro LOCAL_ECONOMY / legend=legend1; legend1 position=(middle right) across=1 mode=share label=("&label" position=top j=c); format LOCAL_ECONOMY comma8.4; run; quit; ... View more

Hi Folks: Can you help me transform data from HAVE to DESIRED OUTPUT structure? I don't need to output it to a data set.  Thanks in advance.    DATA HAVE; INPUT GROUP Desc $ A B; CARDS; 1 Min 1 4 1 P50 2 5 1 Max 3 6 2 Min 3 7 2 P50 4 2 2 Max 5 3 3 Min 7 5 3 P50 6 4 3 Max 8 6 ; I'm not sure whether proc tabulate or report or transpose the best option.  ... View more

Hi Folks:    I'm running many bi-variate models. I'd like to have all variables along with their parameter estimates and p-value in one dataset afterwards. The code posted works. However, if I could produce  t.&var  in the last data step with a common prefix. I could simply concatenate data with common prefix.  Any suggestions to come up with the datasets with a common prefix from output ods in macro?    data t.commonprefix: ; set t.&var._par; if _n_=2 then output; run;   ods trace on; ods output ParameterEstimates=t.&var._par(keep=Parameter Estimate Probt); proc countreg data=alldata Wmat=Wmatrix; model n_covid = &var p55_64 p65_74 P75_84 p85/offset=ln dist=negbin; spatialeffects db_trt_rate p55_64 p65_74 P75_84 p85; run; ods trace off; data t.&var; set t.&var._par; if _n_=2 then output; run; ... View more