Imagine you have a population that is NOT normally distributed. Take a large sample of it, and calculate the mean of your sample. Then take another large sample of that population and calculate its mean. Keep doing this. Think of each sample mean as a data point that gives you an estimate of the mean of the whole population. The Central Limit Theorem says (approximately, and there are conditions) that those sample-mean data points of yours will be normally distributed, even though they were obtained by sampling a non-normal distribution. This also works when the population IS normally distributed, but that's not a surprise. The surprise is that the sample means are normally distributed even when the population isn't. (And again, there are some qualifications I'm not mentioning.)
So, sample means (estimates of the population mean) tend to be normally distributed even when the population they are sampling is not.
The D3 demo here doesn't show multiple sample means, it shows that as you take more and more items from a normal population, the distribution of your sample of items gradually comes to resemble the distribution of the population as a whole.
So, sample means (estimates of the population mean) tend to be normally distributed even when the population they are sampling is not.
The D3 demo here doesn't show multiple sample means, it shows that as you take more and more items from a normal population, the distribution of your sample of items gradually comes to resemble the distribution of the population as a whole.