About sophia_SAS

sophia_SAS · ‎12-02-2015

Hi Rob, Thanks for your reply. You are right. My coordinates were not westlong radians. From some of your previous codes, I found the conversation formula and converted the coordinates from the CBSA map. I then combined the dataset and added a project gproject step. However, the county boundaries aren't showing up on my map - only the CBSA boundaries. I will review the annotate link you sent and see if that helps with why the county boundaries are not showing up. Thanks.

sophia_SAS · ‎12-02-2015

Hi everyone, I'm working on a map that overlays the 2010 Core Based Statistical Areas (CBSAs) with the US county boundaries. I've imported the CBSA shapefiles from the US Census websites and am trying to combine it with the SAS supplied maps.counties dataset. However, when I combine the two I get an error message that says "data system requrested, but value is not on graph 'X', 'Y'. I'm a bit confused about what I have done wrong. When I run proc gmap without the maps.counties overlay. I get the CBSA boundaries. Below is a sample of my code. Thanks! /*Importing ESRI file*/ proc mapimport datafile=usdata out=cbsa10;run; /*restricting CBSA boundaries to continental US*/ data cbsa10b; set cbsa10; state=SUBSTR(NAME10,LENGTH(NAME10)-1,2); if state in ("AK","HI","PR") then delete; run; option mprint; data anno_county; set maps.counties (where=(state not in (02,15,72))); length function color $8; anno_flag=1; xsys='2'; ysys='2'; hsys='3'; when='a'; color='cx00ff00'; size=2.0; if _n_=1 then function='poly'; else function='polycont'; run; data combined; set cbsa10b (drop=state) anno_county; run; data cbsa10b anno_county; set combined; if anno_flag=1 then output anno_county; else output cbsa10b; run; data mydata2; set mydata; if ST in ("AK","HI","PR") then delete; run; proc gmap all data=mydata2 map=cbsa10b all anno=anno_county; id CBSAFP10; choro myvariable / discrete coutline=black; run; quit;

sophia_SAS · ‎11-04-2015

Given that, is there no way to display a map using MSA boundaries via SAS? Thank you for your time!

sophia_SAS · ‎11-04-2015

Hi Darrell, Thanks for your reponse. I think you've got the problem. I've attahed a pdf of my output. As you can see I have a combination of 'blank' (i.e. no color other than outline), colored polygons and lots of empty space. The colored polygons matched the 8 observations from my "test" dataset. However, I thought by using the "all" options on proc gmap, the output would include all values for METDIVFP10, even if they were not included in my 'test' dataset. Does this make sense?

sophia_SAS · ‎11-04-2015

Thank you for your reply! I'm sorry for the copy and paste issue. Please find below the correct 'test' dataset, With that said, when I run the updated datset with the right METDIVFP10 values, I still get the same problem. Below is the updated 'test' dataset. What do you see when you run it? data test; input METDIVFP10 $5. poptest 8.2; datalines; 14484 5.56 15764 11.76 15804 7.01 29404 22.22 31084 5.88 36084 61.20 13644 43.20 19124 88.05 ;; run;

sophia_SAS · ‎11-03-2015

Hi everyone, Using Mapimport, I imported the US Census Metropolitian Division 2010 shp file into SAS. However, when I combine the map file ("metdiv_shp") with my data ("test"), my final map only includes a small fraction of my observations though there should be many more. Here's a sample of my code. The ID variable ("METDIVFP10") for both the data file and the map file are exactly the same. Any thoughts on why my output is limited to a few observations? Am I missing something with my proc gmap code? Or the import command? When I use the SAS supplied maps (i.e. maps.us), it works fine. Thanks! proc mapimport datafile="r:\tl_2010_us_metdiv10.shp" out=metdiv_shp; run; data test; input METDIVFP10 $5. poptest 8.2; datalines; 00080 5.56 00120 11.76 00160 7.01 00200 22.22 00220 75.00 00240 5.88 00320 40.00 00380 83.34 19000 61.20 37200 43.20 ;; run; proc gmap data=test map=metdiv_shp all; id METDIVFP10; choro poptest /discrete coutline=black; run; quit;

sophia_SAS · ‎08-06-2015

Hi Mark! I figured out a solution! Thanks again!

sophia_SAS · ‎08-06-2015

I'm sorry for not mentioning the multiple years of data. I was trying to simplify things . . . As for your question, the only difference in my variables is how the months are labeled. There are actually 36 months with Mon1=Jan 2013 and Mon36=Dec2015 everything else is the same.

sophia_SAS · ‎08-06-2015

Hi Mark! Your solution works great. Thank you for your time! One other follow-up question. I actually have multiple years of data that I'm working with so my start and end dates frequently have different years. For example, subject MM start date= 06/05/2013 and end date = 12/31/2015. Do you have a suggestion on how to work with those differences? Thanks again!

sophia_SAS · ‎08-06-2015

Thanks Reeza and Ballardw for your questions. See my responses to your questions. Let me know if I should clarify further! Thank you! 1. What do mon1 mon2 represent, how are they derived? mon1, mon2, . . . mon12 = the months in 2013 where mon1 = Jan 2013, mon2 = Feb 2013 ... mon12=Dec 2013. 1 = enrolled; . = not enrolled. For example, subject AA was enrolled from mon3-mon12 or from March 2013 to Dec 2013. 2. How do you determine/define continuous enrollment? Continuous enrollment is define by enrollment (i.e. value=1) for at least 1 day in every month from start date to end date or date of event. For example, Subject AA was continuously enrolled for the defined period because there is a 1 from Mon3-Mon12 and AA start date is 3/10/2015 (Mon3) and end date is 12/31/2015 (Mon12). In contrast, Subject CC was not continuously enrolled for the desired time frame because CC had start date and end dates from Jan 2013-Dec 2013 and was not enrolled in mon5-mon9, and mon11-mon12. 3. Provide an example of what you want the output to look like. Ideally, the output will contain at least 2 new variables: eligible (i.e. met the continuously enrollment criteria) and if yes, the sum of months from start to end or start to event. For example, see the same dataset below with two new variables. data want; input SubID startdate eventdate enddate mon1 mon2 mon3 mon4 mon5 mon6 mon7 mon8 mon9 mon10 mon11 mon12 eligible sum_months; datalines; AA 03/10/2015 . 12/31/2015 . . 1 1 1 1 1 1 1 1 1 1 yes 10 BB 05/20/2015 08/02/2015 12/31/2015 . . . 1 1 1 1 1 1 1 1 1 yes 4 CC 01/18/2015 . 12/31/2015 1 1 1 1 . . . . . 1 . . no . DD 08/01/2015 . 12/31/2015 . . . . . . 1 1 . 1 no . EE 02/25/2015 . 05/11/2015 12/31/2015 . 1 1 1 1 1 1 1 1 . . 1 yes 4 FF 04/09/2015 11/10/2015 12/31/2015 . 1 1 1 1 . 1 1 1 1 1 . no . ;;

sophia_SAS · ‎08-06-2015

Hi SAS community, I would like to identify the individuals who have been continuously enrolled beginning from their start date to either end of study (12/31/2015) or date of event. It is possible for a subject to be enrolled prior to the start date. However, I do not want to count the months before the listed start date .Please advise on suggested coding tips. Below is a simplified version of the dataset. Using this example, the Individuals who meet the criteria, include AA, BB, EE. data have; input SubID startdate eventdate enddate mon1 mon2 mon3 mon4 mon5 mon6 mon7 mon8 mon9 mon10 mon11 mon12; datalines; AA 03/10/2015 . 12/31/2015 . . 1 1 1 1 1 1 1 1 1 1 BB 05/20/2015 08/02/2015 12/31/2015 . . . 1 1 1 1 1 1 1 1 1 CC 01/18/2015 . 12/31/2015 1 1 1 1 . . . . . 1 . . DD 08/01/2015 . 12/31/2015 . . . . . . 1 1 . 1 EE 02/25/2015 . 05/11/2015 12/31/2015 . 1 1 1 1 1 1 1 1 . . 1; FF 04/09/2015 11/10/2015 12/31/2015 . 1 1 1 1 . 1 1 1 1 1 . ;; Thanks!

sophia_SAS · ‎04-01-2015

I'm using sgscatter and have two questions about controlling the output. 1. On the output, I would like to include a legend that includes 'fit' and prediction/confidence limits. It would be akin to the legendlabel="Fit" option in sgplot. 2. Is there a good work around to altering my xaxis value? I would like to control these for a couple of my plots, but am unsure of how to do so in sgscatter plot. Is it even possible? Below is the basic structure for my code. proc sgscatter data=have; plot a1*b1 a2*b2 / reg =(clm cli) ; Thanks!

sophia_SAS · ‎03-27-2015

Using proc gmap, how do I should clusters if I only have 1 zip code assigned to each data point? For example, for zip code 99999, I have n=5 data points. This meaning I would like to show 5 dots clustered in the 99999 zip code area. Right now my code only gives me 1 dot per zip code. Thanks!

sophia_SAS · ‎01-23-2015

Hi SAS community, I would like to format the proc format output. How do I get SAS to display the percent values rounded to 1 decimal? Can I add custom formats as well, specifically parentheses around the percent value? For example, ColPer = 55.25. I would like the output to read (55.3). Thanks!eq

sophia_SAS · ‎10-09-2014

Thanks! This worked great and was a lot easier than a macro.

Online Status	Offline
Date Last Visited	‎07-09-2020 05:04 PM

Re: local macro

Re: local macro

Re: local macro

local macro

Re: rolling calculations - duration of admission

rolling calculations - duration of admission

Re: Array and Column Names

Array and Column Names

Re: SAS numeric length

SAS numeric length

Re: Using Macro to create new variables

Re: fill in values in blank excel cells via sas

Re: by variables assigning new variables

Re: new date variables

Re: creating new variables with dates

Re: CBSA and State boundaries gmap

CBSA and State boundaries gmap

Re: proc gmap MSA mapimport

Re: proc gmap MSA mapimport

Re: proc gmap MSA mapimport

proc gmap MSA mapimport

Re: continuous enrollment counting

Re: continuous enrollment counting

Re: continuous enrollment counting

Re: continuous enrollment counting

continuous enrollment counting

sgscatter legend and axis values

proc gmap

proc freq format output

Re: Using Macro to create new variables