Help using Base SAS procedures

sophistocated graphing

Reply
N/A
Posts: 0

sophistocated graphing

I have here an example of an overlay, but I think G plot was used. Can anyone tell me the coding?

In this report:
A. The horizontal lines represent low and high values (range of scores from evaluators)
B. The open circles represent the score of the "self evaluation"
C. The closed (blue) circles represent the "evaluator scores [average]"

I am not able to post the graphic because there is no attachment feature...can I send it to you via e-mail? Please help!
Super Contributor
Super Contributor
Posts: 3,174

Re: sophistocated graphing

You may find suitable comparison graph examples in the SAS-hosted DOC or at various user-maintained SAS-related websites.

Scott Barry
SBBWorks, Inc.

Suggested Google advanced search arguments, this topic / post:

sas graph examples site:sas.com

sas graph examples site:robslink.com

sas graph examples site:lexjansen.com
N/A
Posts: 0

Re: sophistocated graphing

thanks so much
SAS Super FREQ
Posts: 8,744

Re: sophistocated graphing

Hi:
Here's an example of using GPLOT.
http://support.sas.com/kb/25/584.html

It is from one of the SAS/GRAPH sample sites (the type you would find if you searched as suggested by Scott in his post). The GPLOT documentation also has a sample section specific to GPLOT:
http://support.sas.com/documentation/cdl/en/graphref/63022/HTML/default/viewer.htm#gplotchap.htm (has 10 examples at the end of the doc section)

And this public site, by Dr. Allison, has many, many GPLOT (and other examples) which may be useful to you:
http://robslink.com/SAS/Home.htm

If you use GPLOT, you generally want to create a plot from data that will have X and Y values. For example, if you were generating a time-based plot, you might have MONTH on the X axis and WEIGHT on the Y axis; or DAY on the X axis and PRICE on the Y axis. You might add a classifier variable to this type of plot, so, for example, you would have DAY on the X axis and PRICE on the Y axis and different symbols would be used for COMPANY (assuming that you wanted to compare PRICES for each COMPANY for each DAY). Generally with plots of this sort, you might join the plot points so you could see behavior across time (up trend or down trend).

Or, if you were doing a scatter plot, you might have WEIGHT on the X axis and CHOLESTEROL on the Y axis in order to see whether there were any clusters of values for CHOLESTEROL for WEIGHT. You might add a classifier variable to this type of plot, so for example, you would have females indicated with one color or type of marker and males indicated with a different color of type of marker. Generally, with plots of this sort, you might want to see clusters of values or add regression lines or confidence interval lines to the scatter plot.

These are just a few of the plots that you can generate with PROC GPLOT (and if you're using SAS 9.2, you may want to investigate the SG procedures, such as SGPLOT). But without knowing more about your data and the structure of your data, and understanding what you would want to see plotted on the X axis and the Y axis, it is hard to come up with more concrete suggestions.

cynthia
N/A
Posts: 0

Re: sophistocated graphing

Thanks much...it will take me a while to look at all this...very useful!
Super User
Posts: 9,687

Re: sophistocated graphing

Hi.
I think you should post it to Graph forum.
Just as Cynthia said there is a sas-man named Robert Allison who is expert about it,
you can email him,he must can solve your problem.


Ksharp
SAS Super FREQ
Posts: 8,744

Re: sophistocated graphing

I would hope that folks would check out the web site links, documentation and/or post in the SAS/GRAPH forum before emailing anyone directly.

In addition to posting in the forums, Tech Support is another appropriate channel for seeking help, once there is actual code and a meaningful question that can be sent to them. There are many, many examples of series plots and scatter plots in the SAS documentation and in user-group papers and on publicly available web sites.

The way I read the original question, the OP hadn't really tried any code yet and needed to do some more some investigation and research to determine the best method for accomplishing the desired output. Given that the X and Y variables and the structure of the data were not clearly defined in the original post, it is entirely likely that any answer will be nothing more than guesswork based on assumptions. Until some code has been tried, or some more requirements are identified, there's not a lot more for anyone to offer.

cynthia
Ask a Question
Discussion stats
  • 6 replies
  • 187 views
  • 0 likes
  • 4 in conversation