2
$\begingroup$

I don't think this type of plot has a name, but as the legend shows, the shades of gray represents the quantile distribution of repeated temperature measurements, taken on the same calendar day over multiple years, and plotted as time-sequence. It is basically a variation on jamming together a sequence of 92 box-and-whisker plots, placing them each above the date on the x-axis.

enter image description here

I am struggling to quantify and describe the trends that I can see, because this plot combines two types of trend analysis. Time series can use simple language like "rising" and "falling"; distribution plots can use specific words like "skew". But how do I describe the trends in a plot like this where each point on the time-axis has a different spread and skew? For example, there's a spot in late September with a light-gray downward peak spanning several days--how would I describe that peak?

$\endgroup$

closed as unclear what you're asking by Michael Chernick, user158565, Siong Thye Goh, mdewey, Peter Flom May 5 at 13:31

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ Which "mean" and which "population" are you referring to? $\endgroup$ – whuber May 3 at 20:33
  • $\begingroup$ Sorry, I used the word "mean" as shorthand to refer to the time-series of means--I suppose there is one for each day. The word "population" refers to raw data pulled from a weather station. For context, climate change is often predicted as, "3 degrees hotter," and I'm trying to describe what that will mean when it's applied to a continuous distribution. $\endgroup$ – bethyj May 3 at 20:50
  • $\begingroup$ By "raw data pulled from a weather station" what time frame do you intend? A single day, the period shown in the graphic, or a longer period? There's no issue with the interpretation of a mean for a continuous distribution, leaving us to wonder what you're trying to ask. $\endgroup$ – whuber May 3 at 22:50
  • $\begingroup$ Please register &/or merge your accounts (you can find information on how to do this in the My Account section of our help center), then you will be able to edit & comment on your own question. $\endgroup$ – gung May 4 at 4:08
  • $\begingroup$ I don't understand what you are asking. You describe the plot with words (as you have done). One big advantage of a good graph is that it displays a lot of information. So, really, your graph is your description. $\endgroup$ – Peter Flom May 5 at 13:31
2
$\begingroup$

[First, your time axis is confusing (at least for Europeans), I take that it covers july, august, september)?] This is a kind of time-series plot of marginal distributions, not of an actual time series, so it gives no information about time dependence, autocorrelations, and such, which also could be useful in describing climate. So any discussion of this plot should make that clear. So the plot tells you how the median temperature, quartiles, and so on varies with calendar time. Some of the short-time variability in the plot is probably caused by noise, and do not make sense as climate description, so a better plot would use some smoothing.

From the plot we can visually compute a measure of variability such as inter-quartile range, which is remarkably stable. That seems interesting to comment. You could describe in words how the median varies with time, how the quartiles vary with time, ..., and so educate your readers in reading the plot.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.