Programming the statistical procedures from SAS

Which Graph is more effective for data look..

Posts: 44

Which Graph is more effective for data look..

Hi All,

Could anyone let me know that which graph is more effective to see the nature of the data.

client needed on query that "A rough plot of baseline labtest by raw change at week12 with data look subset".

so please clarify it ASAP. Thanks in advance!

Best Regards & Many Thanks,


Super User
Posts: 5,308

Re: Which Graph is more effective for data look..

Try to clarify with your client what the purpose is, in more detail, then it might be easier to chose the appropriate visualization.

Data never sleeps
Posts: 44

Re: Which Graph is more effective for data look..

Hi Linus,

i got the following comment form client please look on it!

The relationship based on Labtest data is complicated!  Would like to have a data driven dialogue with XXX(Company Name) statisticians.  Suggest starting with a rough plot of baseline Labtest by raw change at Week 12 with data look subset. 



Trusted Advisor
Posts: 2,114

Re: Which Graph is more effective for data look..

Reading between the lines, there are a lot of different lab tests and the client is trying to see where there might be some action.  I am assuming the lab tests all have continuous measures, but are on different scales.

There are a number of different ways to attack this.  A number of the SG procedures and ODS graphics could be used.  Look through

and you could find some examples to show your manager to get better direction on the presentation for the client.

One possibility is to look at %-change from baseline.  You could then label each lab test with a bar to represent the percent change; this if often done as a horizontal bar chart.  If you have 'normal values', you could also flag the test that were outside of the normal range at baseline (different color bars, for example).  You could also order the bars by height so the client can focus in on the ones that are most extreme.

Ask a Question
Discussion stats
  • 3 replies
  • 1 like
  • 3 in conversation