BookmarkSubscribeSubscribe to RSS Feed

Pareto plots

Hi Guys,

I have a JMP data table which has multiple tests (100+)in columns and next to each test column there is a column that says of that row (data point) passed or failed that test.

I would like to see which are my top failing tests. Is there a way to do this? I tried using the pareto plots but could not get it to do what I want. Can I use the pass/fail information that is near the test column to sort my top failing tests?

The file looks something like this: (there are 100+ values; I have shown only two values)

Value1 P/F Value2 P/F
250 P 1.56 P
222 P 1.89 P
223 P 1.25 P
21 F 1.58 P
20 F 2.22 F

So I want to see a pareto plot which shows the top fail as being Value1 with the bar on the chart showing 40% followed by another bar on the same pareto plot showing 10% for Value2.

How can I do this?


Super User


Jun 23, 2011

As I understand, a pareto plot shows the relative frequency in a cumulative way, up to 100%. I don't know how you can to use a pareto to plot the fail percentages of 100+ columns. I would easily become a lot more than 100%. Do you mean sorted bar chart?

If so, you can try the script below. The script assumes that your table is organized as in your example (Value1, P/F 1, Value2, P/F 2,...). It can probably be simplified.

Thanks a lot, that worked.
One more question:

What if I have blank cell values? It seems to count the blank cell value in the total.

Can I omit the blank cells?


Super User


Jun 23, 2011

There are probably several ways to do it.


should eliminate any impact of empty cells in the B/F column, although if there are no empty cells it will give wrong number. You need to put those alternatives within a conditional statement.

Something like
if(a[1]=="F", colF = c[1] / Sum( c ) , colF = c[2] / (Sum( c )-c[1]))

Try! I have not time to test now.
Update: As above colF is really followed by the counter i within square brackets althoughthe browser refuses to show that.

Message was edited by: MS

Message was edited by: MS