Chart of the day

[organofcorti]

I love this chart, because before I looked it up I had no idea what it was about:



Don't tineye it. Just look and imagine.....

I'll go first: A competition. For stick carrying. Carried out over a whole year. Clearly the younger competitors are much better at carrying sticks than the older ones, probably because the older ones have better things to do than carry sticks. Edit: probably iPhone owners having sex

The dotted line is probably girls. They stop doing stupid things sooner.

Better suggestions?

[ZodiacDragon84]

I love this chart, because before I looked it up I had no idea what it was about:



Don't tineye it. Just look and imagine.....

I'll go first: A competition. For stick carrying. Carried out over a whole year. Clearly the younger competitors are much better at carrying sticks than the older ones, probably because the older ones have better things to do than carry sticks. Edit: probably iPhone owners having sex

The dotted line is probably girls. They stop doing stupid things sooner.

Better suggestions?

Amount of Time Fido carries a stick?

[organofcorti]

Well, that would explain the incidence of stick carrying in the over 20 year old age group. But a peak of 1.6 stick carrying episode per 1000 hours? On average, that would mean one stick every 26 days and that the kids weren't doing their job. No, i'd hate to think that Fido was only getting to the park once a month.

[bb113]

I love this chart, because before I looked it up I had no idea what it was about:



Don't tineye it. Just look and imagine.....

I'll go first: A competition. For stick carrying. Carried out over a whole year. Clearly the younger competitors are much better at carrying sticks than the older ones, probably because the older ones have better things to do than carry sticks. Edit: probably iPhone owners having sex

The dotted line is probably girls. They stop doing stupid things sooner.

Better suggestions?

Amount of Time Fido carries a stick?

The x-axis implies that each data point is the mean of some sample. Also the data range is only 16 years long while the age range goes out to 35, so I don't think this was a naturalistic study (there would be problems in gathering the age of subjects older than 16 years). So it is a longitudinal study of some captive or domesticated population. "Carrying" implies it is an animal, unless "stick" means something weird.

What the hell carries 1 stick every 26 days one year and then never carries a stick the next? Maybe stick carrying is dangerous, and usually kills you by year 9. Or, probably more likely, stick carrying is part of a mating ritual. The animal (a bird maybe), needs to try harder and harder to mate each year until most die from natural or stick-carrying causes by year 9. After that only the ones that don't bother carrying sticks any more survive until old age.

[organofcorti]

I love this chart, because before I looked it up I had no idea what it was about:



Don't tineye it. Just look and imagine.....

I'll go first: A competition. For stick carrying. Carried out over a whole year. Clearly the younger competitors are much better at carrying sticks than the older ones, probably because the older ones have better things to do than carry sticks. Edit: probably iPhone owners having sex

The dotted line is probably girls. They stop doing stupid things sooner.

Better suggestions?

Amount of Time Fido carries a stick?

The x-axis implies that each data point is the mean of some sample. Also the data range is only 16 years long while the age range goes out to 35, so I don't think this was a naturalistic study (there would be problems in gathering the age of subjects older than 16 years). So it is a longitudinal study of some captive or domesticated population. "Carrying" implies it is an animal, unless "stick" means something weird.

What the hell carries 1 stick every 26 days one year and then never carries a stick the next? Maybe stick carrying is dangerous, and usually kills you by year 9. Or, probably more likely, stick carrying is part of a mating ritual. The animal (a bird maybe), needs to try harder and harder to mate each year until most die from natural or stick-carrying causes by year 9. After that only the ones that don't bother carrying sticks any more survive until old age.

This is definitely my favourite theory so far. A domesticated population killing themselves because mating demands random incidences of dangerous stick-carrying followed by long periods of boredom. It sounds like a definition of modern warfare.

[bb113]

Crap, I meant 13 years, not 16...

Also, 1.6 sticks per 1000 hours doesn't necessarily mean 1 stick every 26 days. It could just as well be 14 sticks one day per year. It would be strange to bin by 1000 hours in that case though. However, sometimes people like to normalize so their charts center around 1, and if my interpretation of "mean age" is correct this chart is missing error bars already, so there would be no point in complaining that it doesn't show the underlying distribution.

26 six days is about a month. So perhaps this is something that happens according to lunar cycles.

This is definitely my favourite theory so far. A domesticated population killing themselves because mating demands random incidences of dangerous stick-carrying followed by long periods of boredom. It sounds like a definition of modern warfare.

 

[organofcorti]

Crap, I meant 13 years, not 16...

Also, 1.6 sticks per 1000 hours doesn't necessarily mean 1 stick every 26 days. It could just as well be 14 sticks one day per year. It would be strange to bin by 1000 hours in that case though. However, sometimes people like to normalize so their charts center around 1, and if my interpretation of "mean age" is correct this chart is missing error bars already, so there would be no point in complaining that it doesn't show the underlying distribution.

Well, I'd been assuming it was a daily stick carry of 0.0384 of a sticks per day initially. Then I remembered what my mum would say on long car trips when I asked how much longer the journey would take. She'd tell me "How long is a piece of stick?" Ah, mothers' wisdom. Always useful.

26 six days is about a month. So perhaps this is something that happens according to lunar cycles.

Back to thinking this might be a human population? I know of someone who'd like to carry a big stick and have at me with it, about every 26 days.

You're so close I'll give you a few hints:
You're correct about it being a captive population, but they are not domesticated
From the text:

To control for age and satisfy small samples of individuals per age category, each individual was assigned to one of 11 age-classes. Assignation to age-class was determined by the individual's mean age between their start and end of observations, 1993-2006

The text also mentions sexual swelling (nudge nudge, eh?)
Mean stick-carrying rates across individuals were higher for females than for males.


You realise you're using a miniature version of my old avatar for your avatar?

[bb113]

hmm. So animals in the "9 years old" group would have been about 3 years old when the study began and 15 when it ended, and the age was binned in two year intervals. It looks like new animals can be born into the population under study. From this we see the peak corresponds to animals that were 1 years old at the beginning of the study. So the stick carrying is only really occurring in animals that were less than 1 year old at the beginning and increases at least to the age of 13. Animals that were over 1 year old at the beginning do not stick carry. Perhaps there is a critical period (1-2 years old) during which the stick carrying behavior must be learned, and this behavior was then learned by newly born animals entering the study, but does not really manifest itself fully until the teen years (sexual maturity).

I say it was a study of gender identity in primates.

[organofcorti]

I say it was a study of gender identity in primates.

Well done!

Original paper here

And this is your prize!

[bb113]

Short paper

[stochastic]

I refuse to read charts without error bars.

[bb113]

Nice chart:

[organofcorti]

Nice chart? Now I understand why you think your chart is prettier than mine

[bb113]

I have no idea how this thing is made but it sounds awesome:

Here's the caption:

Fig. 1. Approximation of data used for constructing the quality of life index. Each of 192 points represents a country in 4-dimensional space formed by the values of 4 indicators (gross product, life expectancy, infant mortality, tuberculosis incidence). Different forms and colors correspond to various geographical locations. Red line represents the principal curve, approximating the dataset. Three arrows show the basis of principal components. The best linear index (first principal component) explains 76% of variation, while the non-linear index (principal curve) explains 93% of variation.

http://arxiv.org/pdf/1001.1122v2.pdf

And an overview of the method:

Elastic maps provide a tool for nonlinear dimensionality reduction. By their construction, they are system of elastic springs embedded in the data space[1]. This system approximates a low-dimensional manifold. The elastic coefficients of this system allow the switch from completely unstructured k-means clustering (zero elasticity) to the estimators located closely to linear PCA manifolds (for high bending and low stretching modules). With some intermediate values of the elasticity coefficients, this system effectively approximates non-linear principal manifolds. This approach is based on a mechanical analogy between principal manifolds, that are passing through "the middle" of data distribution, and elastic membranes and plates. The method was developed by A.N. Gorban, A.Y. Zinovyev and A.A. Pitenko in 1996–1998.

http://en.wikipedia.org/wiki/Elastic_map

[organofcorti]

The ideas are pretty interesting.

My gripe is that the chart doesn't make itself very clear on a first look. It's ugly and confused. Too many overlays, too many different dimensions, strangely sized and placed legends. Charts are for communication, and if you already need to understand the data before you understand the chart, you're not presenting it the right way.

[bb113]

Well the reason I came across this is I am trying to find a way to visualize the correlation between 6 variables. If you know an easy to understand way of doing this let me know.

Dependent variable: A score between 0-20

Independent Variable 1: A percentage loss of brain tissue
IV 2: X position at which IV 1 is measured (+7 to -3)
IV 3: Y position at which IV 1 is measured (0 to +5)
IV 4: One of three categories (treatment groups)
IV 5: Neuron complexity at each of the locations in the opposite hemisphere of the brain (lets say a single number from 0-60)

The best I've come up with so far is a heatmap like this (I would overlay it onto a picture of a brain):


The values are the correlations (Pearson's R) between IV1 and the DV at each position. So a large IV1 magnitude at position 1 lateral and 4.2 anterior appears to be the most important factor in determining the score. Just ignore the positive correlations, they are due to small sample size and the fact that the tissue is lost roughly in the shape of a 3 mm diameter circle (so loss at 4 lateral means there is little loss at 1 lateral for that individual, etc)

[organofcorti]

I think a heat map is good for the data you've presented. ggplot2 can create very nice heat maps like the one below.



But didn't you mention 5 IVs?

[bb113]

Well the way I did it above each xy area is a separate IV and the DV and IV1 magnitude information is lost.

[organofcorti]

Well the way I did it above each xy area is a separate IV and the DV and IV1 magnitude information is lost.

Ah yes, so you said.

You could try using lattice plots to make the heat maps easier to relate to each other. Like this:

[bb113]

Unless I am misunderstanding something, the lattice plot doesn't really work for me since I wish to show how the different variables interact.