You could refer to this post: ViSta: The Visual Statistics System

What I learn from the post are:

1.Sample Distributionderives from calculation/statistics using the original sample values. See Section 4 in the post.

2. from the figure of Sample Distribution, you firstspeculate/determine which type of distribution it followsby observing the shape of the figure/curve;

3. say it nearly follows normal distribution, then you use the original sample values to derive mean and standard deviation, twokey parameters to model the density functionof normal distribution;

4.Sampling Distributionis then the distribution you just modeled in step 3. See Section 13 in the post.