Best way to display data graphically

[Anita_n]

Dear all,

if I have a data that looks like this.

data have;
infile datalines;
input num_of_pat  drug $12. dosis ;
datalines;
1 Cefalexin    120
1 Calcipotriol 165
1 Folic acid   260
1 Piriton      300
1 Fentanyl     360
1 Folic acid   260
1 Co-beneldopa 356
1 Folic acid   267
2 Cefalexin    119
2 Fentanyl     354
1 Allopurinol  331
1 Folic acid   278
1 Cefalexin    113
1 Allopurinol  332
;
run;

proc sgplot data=have;
series x=dosis y=num_of_pat /group=drug;
run;

what is the best way to display this graphically: The problem is I have over 100 records and the num of patients ranges from 1 to 2 which makes it not easy to display this. Any idea?

[PaigeMiller]

What message do you want to the graph to provide the viewer?

[Anita_n]

The number of patients who received a certain drug and the dosis they received

[PaigeMiller]

I am thinking a clustered bar chart (VBAR) would make more sense to show the variable DOSIS. Horizontal axis is drug, with separate bars for NUM_OF_PAT 1 and 2.

[Anita_n]

Do you mean something like this

proc sgplot data=have;
vbar dosis /  response=num_of_pat group=drug groupdisplay=cluster;
run;

[PaigeMiller · Accepted solution]

proc sgplot data=have;
vbar drug/  response=dosis group=num_of_pat groupdisplay=cluster;
run;

[Anita_n]

Thank you, I think I can go with that.

[Anita_n]

@PaigeMiller  why are the bars summed up per group. For example Cefalexin  occurs three times. I don't want the 3 values summed up but displayed side by side

[PaigeMiller]

So you want two bars for Cefalexin when num_of_pat=1 and one more bar when num_of_pat=2?

[Anita_n]

Yes please

[PaigeMiller]

I don't know how to do that. I don't even think it is possible.

[Anita_n]

okay, thankyou

[PaigeMiller]

Maybe you want the average to be the height of the bar instead of the sum?

[Anita_n]

No, not the average

[ballardw]

I am not clear what the graph is supposed to explain or clarify. What does "dosis" measure?

This will draw a separate needle for each observation.

proc sgplot data=have;
 needle x=dosis y=num_of_pat /group=drug ;
 yaxis values= (0 to 3 by 1);
run;

Should perhaps some  values, such as your Allopurinol with Dosis = 331 and 332, really should be treated separately. But with a purpose for what this graph should explain...