Opposing Bullet Chart Feedback

[Ed Burke]

Hello again,

I'm attempting to create a bullet chart that will allow our leadership to visulize the current state of our program's execution. Attached is a sample bullet. I'm looking for feedback on positives and negatives so I can improve the product. Everything on the chart is based on input provided on a daily basis and there are no static elements (i.e. every elements changes with every time period (background size, bar length, goal bars, requirements markers, etc.).

Here's how to interpret it:
- The background has two different colors (dark grey/light grey). The combination of the two (entire length of the bar) represents the number of events we have scheduled. The light grey represents the number of events we expect to cancel based on historical data. The dark grey represent the number of events we expect to occur as scheduled, based on historical data.

- The black bar (from the left) represents the number of events that have already occurred for this time period (usually a month). The black goal bar (error bar) represents the number of events we expect to have occurred by this time during the time period (i.e. if we're half way through the time period, the goal bar will appear at the half way point).

- The red bar (from the right) represents the number of events that have been cancelled for this time period. The red goal bar represents the number of events we expect to have been cancelled at this point in the time period. Under ideal conditions, the two bars (red/black) will meet at the dark grey/light grey background transition point at the end of the time period. FYI, there will be a certain number of events cancelled. We're not trying to stop that, we're simply trying to control it to keep cancellations at or below historic levels.

- The arrow pointers at the bottom represent various requirements of our programs. For example, the triangular marker labelled "C" represents the number of events required to achieve a specific goal. Same for the marker labelled "R." These labels will change color (Black, Red, Yellow, Green) based on the results we expect to see at the end of the time period (expected results are based on actual rates, tracked by day).

This setup made sense to me because the values of the bars compared to their respective backgrounds and the various requirements markers allow us to analyze where we are in any given time period and identify weaknesses in certain processes so we can meet our goals by the end of the time period.

So ... is it too complicated or too simple? Too difficult to understand quickly? Too ugly? Does it need to be split into two graphs/charts? Any feedback (positive or negative) is greatly appreciated.

Thanks,
Ed

P.S. the values of the individual elements have no real meaning for our analysis. The only important value on the chart is the total number of scheduled events (i.e. total length of both background together). This is what the x-axis value represents.

 

[Luke M]

Hi Ed,
First, I like the effort and style of the chart. Nice combination of different aspects. So, I know my next set of comments is easy to say, but harder to implement.

The problem I see with the chart is that, as you said, the values can't be read easily or at all for the red bar. Also, there's a potential problem if the actual bars exceed expectations...do the bars suddenly overlap? Grow into each other? I'm not sure.
What I would propose is a tornado type effect (see images here for example). Let's your bars expand, and the values are now useful.
This is where might decide to split into two charts. Not knowing your business, I can guess that the two sets of data have a correlation. Only if there is no correlation, than they should be separate charts.

As for clarity of the chart, once it's explained once, I understand it. Might be worthwhile to have a legend for the 2 small points, and some simply data labels on the bars themselves.

Anyways, that's my two cents.

 

[Ed Burke]

Hi Ed,
First, I like the effort and style of the chart. Nice combination of different aspects. So, I know my next set of comments is easy to say, but harder to implement.

The problem I see with the chart is that, as you said, the values can't be read easily or at all for the red bar. Also, there's a potential problem if the actual bars exceed expectations...do the bars suddenly overlap? Grow into each other? I'm not sure.
What I would propose is a tornado type effect (see images here for example). Let's your bars expand, and the values are now useful.
This is where might decide to split into two charts. Not knowing your business, I can guess that the two sets of data have a correlation. Only if there is no correlation, than they should be separate charts.

As for clarity of the chart, once it's explained once, I understand it. Might be worthwhile to have a legend for the 2 small points, and some simply data labels on the bars themselves.

Anyways, that's my two cents.

Luke,

Thanks for the feedback. I considered the tornado chart (I may still create to display an entire year's worth of data, thanks for the link), but we wanted to keep the chart much smaller as we only look at one month at a time in our decision meetings and, more importantly, the actual value of the red hashed bar doesn't really mean anything as far as analysis. The important part is where the hashed bar is in relation to the goal bar and light grey/dark grey background transition point as well as the requirements markers.

The bars can eventually touch, but will never overlap each other. The center section the bullet chart is made up of six values: Events Completed, Gap Filler (Events Completed to Transition Point), Attrition (Red Hash Bar) Overflow, Event Completed Overflow, Gap Filler (Attrition to Transition Point), Attrition. The combination of the two bars is the total of events conducted and events lost (attrition). They can never add up to more than what was scheduled, so the bars will never overlap. The Gap Filler and Overflow values are calculated to ensure the bars display the correct length (in many cases, these numbers are simply 0).

The data is heavily correlated. Too much attrition may keep us from meeting our end-of-month goals ("C" and "R"). Too little attrition, and we'll complete too many events, with the overage taken out of another month because we are only funded to conduct a certain number of events every year (and we cannot go over). If we constantly take out the overage, we affect our training. It's a delicate balancing act.

Again, thanks. I'll look into your suggestions and see about improving the data display.

Ed

 

[NARAYANK991]

Hi Ed ,

I have just one comment ; will those for whom this visualization is meant understand it without all the explanation that you have given for our benefit ?

If yes , then it is a good visualization , if not , then you need to see what is to be done to improve it.

Any visualization is supposed to be self-explanatory , at least to those who have the domain knowledge ; a non-finance person may not be able to understand a visualization which to a finance person might be immediately understandable.

Narayan

 

[Ed Burke]

Narayan,

It is not fully self-explanatory. With a couple of labels, though, anyone within the decision making group at our company could easily tell what the information means, though they may need an explanation of the goal bars and reasoning for the colors of the requirements markers since they aren't used to seeing bullet charts.