Google searches yield the typical solution for side-stepping this plotting issue — three different sized circles are layered on top of one another in powerpoint with the broad bands named, but the placement of the individual questions that form the foundation for each of the bands is hidden. Here is the only example I’ve found with the actual points plotted, albeit for a single ring: http://www.psy.gla.ac.uk/~frank/Talk_folder/Y2K_Talk_folder/sld019.htm

Your radar chart of oil prices reminded me of this graphical struggle I’ve been in. I am reaching out to see if you had dealt with the circumplex before.

In fact I did notice the bullet points on your labels, but it was not clear (assumed maybe) that the bullets marked the time of the event, or they may have just been bullets. The span of the labels doesn’t matter much (one of yours reached across nearly five years) as long as the time it references is clear. Perhaps the vertical lines on my chart are a bit more precise than bullet points, especially since they used the actual dates of the events to position the lines. In any case, if my labels require appended dates to clarify when they occurred, then so do yours. (I think neither requires this, as long as the bullets can be relied upon to pinpoint the dates.)

The initials of each month were plainly obvious on your chart. That’s not a cause for confusion. Making the spokes proportional in length to the number of days in each month was an interesting feature, but it was merely decoration. The data I show has greater resolution than your monthly bars, so it stands to reason that the position of the label callouts has this same fine resolution. Even if a viewer didn’t know the data was weekly, s/he would see that it was more granular than monthly. The Vegas flood occurred at the beginning of July, and given the resolution of my chart’s plotted data, you can judge that the line occurs somewhere at the end of June or beginning of July. More important is pointing to the precise point on the line chart, to see whether there may be an effect. In fact, I thought it was a strange event to plot: a priori I could not imagine why a flood in Vegas would have the slightest effect on fuel prices, and I don’t think the data shows any effect.

You failed to address the largest flaw in your chart, the angular misorientation of the bars in your chart which makes comparison of different months inaccurate if they are separated by more than a few intervening months. The second largest problem is the rotational symmetry which falsely implies a cyclical nature to the information it tries to convey.

I don’t understand the problem with publishing my chart, unless it’s just not stylish enough. I did state the purpose of my chart, though, which was to make a chart which is easier to interpret and more clearly labeled.

Sorry for my late reply. There are a couple of things that I wanted to correct from your post. For example when you say that ‘it’s hard to tell which dates the labels reference’ I would advise you to look closer because there’s a bullet in the beginning of each one. As for the ‘some of the wrap far enough to span up to two years’ your go as far as span for 4 years.
On the other hand, when you placed some text in blue ‘which I repositioned to more accurately indicate their occurrence’ you’ve really made things a lot harder. You might not understand it; but each of the years on my chart have the first letter of the month on the bars. For example, in 1999 the ‘major flash flood in Vegas’ event happened when on your chart? Somewhere in the middle of the year, as in my chart you can see that, not only it happened in July, but also at the beginning of the month (therefor it’s probable that this event influenced prices on the same month, as opposed to events at the end of a month that would only influence the coming month).
And, as much of the discussions that I’ve read between ‘excel’ charts and some trials from visual designers, there is really one point that is always overlooked. And this shouldn’t be like that. In my professional experience (I work on a newsmagazine) your chart would be impossible to publish. It might work wonders on a powerpoint presentation for a corporation (much better then mine I give you that) but it wouldn’t work for a printed media. Let’s assume that you would use CMYK colors, you would still have your strokes to much alike, to much text justified to the right. You have a lot of labels that are ‘glued’ together, no clear indication of when the events of the labels took place (you might try adding the actual day and month in braces after the label) and still have a very large graph (even if it is smaller in height then mine) with the problem that, since the strokes are to thin you would loose some information if you had to print the chart on a spread page to actually be able to read something.
My conclusion? Besides the obvious point that I don’t agree with most of what you posted, it would really help this kind of discussions if we made clear from the start what the graphs that we make will be used for.
My best

When you start with an appropriate chart type, the rest of the features you mention can be added with a few dropdowns or checkboxes, with a few worksheet formulas.