I found this link a bit helpful. It gives a basic over-view but falls short of a thorough understanding of things such as why prime, why bit shift etc..
From this link, highlight some section that gave an overview
What Makes A Good Hash Function
Most good hashing functions work by computing the remainder after dividing by the table size N.
This always gives a value between 0 and N-1 so it suitable but if N is a prime number then it is also excellent at scattering the data round the table. Of course, if you have a text value that you want to hash you first have to convert it into a suitable numeric value and a simple scheme like the one in the example will not do.
You need to produce a different numeric value for every possible text value and adding together the ASCII codes of the first two letters clearly doesn't work. A better method is to weight each of the ASCII codes by the position of the letter by multiplying by 1 for the first character, 10 for the second, 100 for the third and so on.. before adding them up to give a single value.
In general building a really good hash function is difficult and in most cases you need to find one that has good properties and has been well tested.