I am incredibly proud of a chart I've just done

[Lux Interior]

Lemon? Looks like puss.

[Greg]

space-time wrote:
come on just change the company name, modify a few numbers and post the damn chart.

We need to know if you're proud of a masterpiece or just a run-of-the-mill Excel chart.

http://files.me.com/gdermer/1xjpr4

I'm very open to feedback on this one; fire away.

[ScottG]

So the certain thing must happen at least once every 90 days? Is that correct? This data must have been collected over a fairly wide time period. The only thing that I would try that would be different would be to have two y-axis, one for the blue columns and one for the cumulative %, rather than having the actual values on each column. Whether this would actually be an improvement, I couldn't say.

If I interpreted the graph correctly, then yes, I like it.

cheers

scott

[Greg]

ScottG: thanks for the feedback!

I gotta run, but will post a bit more explanatory stuff later...why those numbers are there, etc.

[deckeda]

Should the chart include value judgments or conclusions about the number of times? Looking at the chart, I'm reading that the thing must happen but don't know if more=better.

At first glance it looks like if the thing occurred fewer than 4 times, and especially fewer than 2 times, that the rules were much better adhered to. But I don't know if that's a goal. So is the cumulative percentage a known-goal by whomever needs to see the chart, or would the chart do better to directly state the goal(s) as well somehow?

[Mike Sellers]

Yep, that is a damn fine chart. Damn fine.

[Greg]

ScottG wrote:
So the certain thing must happen at least once every 90 days? Is that correct? This data must have been collected over a fairly wide time period.

Sorry; that wasn't clear.

The idea is kind of like if you had a collection of pancakes, each of which needed to be flipped every 90 seconds. If that's what the chart represented (substituting "seconds" in our brains for the "days" that's on the chart), then you could see that 6 pancakes were flipped between 0 and 10 seconds ago, 8 pancakes were flipped between 10 and 20 seconds ago, and so on. It's capturing data for a large group at a point in time.

ScottG wrote:
The only thing that I would try that would be different would be to have two y-axis, one for the blue columns and one for the cumulative %, rather than having the actual values on each column.

What you've described is certainly common, as that's the default arrangement in Excel. The problem with this is what the labels here are designed to overcome. That is, we're all familiar with someone looking at one of the columns, trying to figure out what value it represents, and tracing their finger across to the left hand column. The individual labels create more ink on the chart, but they also simplify this particular task.

Labels are left off of the cumulative percentage line, because it's only real purpose is to show that 71% of the pancakes have been flipped appropriately.

In other words, there's a fairly significant chance that someone will ask "How many pancakes were flipped between 190-200 seconds ago?" Therefore, the columns are labeled. Not so for the red line, so it is only labeled once.

[Greg]

deckeda wrote:
Should the chart include value judgments or conclusions about the number of times? Looking at the chart, I'm reading that the thing must happen but don't know if more=better.

At first glance it looks like if the thing occurred fewer than 4 times, and especially fewer than 2 times, that the rules were much better adhered to. But I don't know if that's a goal. So is the cumulative percentage a known-goal by whomever needs to see the chart, or would the chart do better to directly state the goal(s) as well somehow?

deckeda,

Not so much about the number of times something occurred; rather about how many of a group are meeting a particular threshold, right now. Please tell me whether that clarifies.

[Greg]

Excel geekery:

By default, once labels are applied to the columns, all zero-height columns are tagged with "0". What's the fastest way to the program hide these zeroes?

Note: I know that, grammatically, it's "0." However, I don't care. :-)

[deckeda]

Greg wrote:
[quote=deckeda]
Should the chart include value judgments or conclusions about the number of times? Looking at the chart, I'm reading that the thing must happen but don't know if more=better.

At first glance it looks like if the thing occurred fewer than 4 times, and especially fewer than 2 times, that the rules were much better adhered to. But I don't know if that's a goal. So is the cumulative percentage a known-goal by whomever needs to see the chart, or would the chart do better to directly state the goal(s) as well somehow?

deckeda,

Not so much about the number of times something occurred; rather about how many of a group are meeting a particular threshold, right now. Please tell me whether that clarifies.
It does.