Regression plot using GPLOT

[TMAlves]

Hi there,

I have the following script to plot regression models, but there are four main problems that I could not figure out using the available manuals:

1)    Remove minor ticks in both axis.

2)    Include the linear regression model inside the graph using three decimal cases and rename the variables, i.e., Y = a – bx (R2 = 0.003).

3)    Graph frame should be thicker.

4)    Size of the number of both axis should size 12pt.

I included the current output and the desired graph to ilustrate what I need.

Thank you so much for considering helping.

.

LEGEND1 value=(f=times h=12pt)

POSITION= (bottom center inside)

MODE=protect

noframe;

AXIS1 major =(h=1)

order=(0 to 3600 by 400)

LABEL=(ANGLE=360 ROTATE=0 HEIGHT=14pt font=times 'Cumulative aphid-days');

SYMBOL1 pointlabel = ("#Plot" h=2 font=times h=10pt color=black)  value=none I=RL;

AXIS2 major =(h=1)

order=(0 to 1 by 0.2)

LABEL=(ANGLE=90 ROTATE=0 HEIGHT=14pt font=times 'NDVI');

SYMBOL2 color=black font=times;

PROC SORT DATA=ALLOPENP; BY DateA Place; RUN;

proc gplot data=ALLOPENP; BY DateA Place;

plot NDVI*CUMADAYTaphid/

haxis =axis1 

vaxis =axis2

caxis=black

regeqn

legend=legend1;

run;quit;

[GraphGuy]

1)    Remove minor ticks in both axis.

axis2 minor=none;

2)    Include the linear regression model inside the graph using three decimal cases and rename the variables, i.e., Y = a – bx (R2 = 0.003).

You'll probably have to calculate the values, and then stuff them into a custom text string, and annotate it onto the graph.

3)    Graph frame should be thicker.

axis2 width=2

4)    Size of the number of both axis should size 12pt.

axis2 value=(height=12pt)

[TMAlves]

Thank you Robert.

Considering your suggestion for question (2), I calculated the values and called them in a macro variable.

However, I have a BY statement and the macro variable saved only the last intercept and slope.

Any idea how to save the values for each DATE and PLACE? I am thinking to use each value in a custom text string onto the graph.

PROC REG DATA=ALLOPENP; BY DateA Place;

MODEL NDVI=CUMADAYTaphid;

ODS OUTPUT ParameterEstimates=estimates FitStatistics=stats;

run;

DATA reg_results; MERGE estimates stats; BY DateA Place;

    if variable="CUMADAYTaphid" then call symput("slope", estimate);

    if variable="Intercept" then call symput("intercept", estimate);

    if Label2="R-Square" then call symput("rsquare", cValue2);

run;

[Reeza]

You need to add in an N to your slope, so that you have multiple slope macro variables, ie

slope1 slope2 ... slopeN

DATA reg_results; MERGE estimates stats; BY DateA Place;

    if variable="CUMADAYTaphid" then call symput("slope"||put(_n_,2. -l), estimate);

    if variable="Intercept" then call symput("intercept"||put(_n_,2. -l), estimate);

    if Label2="R-Square" then call symput("rsquare"||put(_n_,2. -l), cValue2);

run;

[KenDodds]

For putting a regression equation on each graph, have a look at the ANNOTATE facility. You can use that with BY group processing, so once you have figured out how to use it for one graph (takes a bit of work) it should be straight foward getting what you want for all the graphs.