11-26-2015 04:51 AM
I am new to SAS Univerity edition. I am trying to learn SAS as much as possible for my PhD studies. However, I am not able to produce the graph i have attached with the file with SAS UE. It has 3 different y axis on left side. I can use Y2AXIS to use right hand side y axis, but, in case i need to draw 4 or 5 different Y axis, what will be the solution? The plot is drawn with R. I have previously used R and Veusz plotting library. I didn't find any solution for this with ODS graphics. SAS/GRAPH is absent in university edition.
11-26-2015 05:26 AM
Well, to be honest, I am not sure its possible. Never seen a graph with three differing axis on one side before. There is a very good resource at:
But a breif search and I couldn't see anything there. Now its possible to force it somewhat, i.e. create one graph with Number of Groups, and have x axis have 9000, 8000 as ticks on X axis but not displayed. Then overlay a line plot which has all the same X value, say 9000 for Median Group Size, and draw a line up with data points labeled to get the effect of having three Y axis. Not sure if thats makes sense, and it would be a fair bit of work. From looking at the figure presented however, it seems over complicated and quite confusing, e.g. Groups=6, Population=600, Median groups = 1.5 (roughly). Would you not be better off reversing the axis, so put cal BP on the Y Axis. Then on the X Axis have your multipe information as Sanjay demonstrates in this post:
11-26-2015 10:56 PM
You can get equally informative comparative graphs for any reasonable number of variables with COMPARE graphs in proc sgscatter. BTW, experience tells me that overlying time series on different scales can sometimes give a false impression of a strong relationship between the variables.
11-27-2015 12:38 AM
This type of graph is quite common in fermentation technology where it is used to present simulation of variables like product formation, substrate uptake, ph/ compound variation during fermentation period. Once this kind of plot is pubished by some reputed journal, others are just started copying and dosenot take other types of visualization, even if the former is incorrect.