So a fairly cool paper is currently in preprint by some Stanford Mathematicians Kannan Soundararajan and Robert Lemke Oliver- regarding the unexpected distribution of consecutive primes.

**Spoiler alert:** *the last digits of consecutive prime pairings aren't randomly distributed*.

Wait what. What does that even mean? It means if (for example in base 10) I have a prime $p_n$ that ends in the digit 1, there's a high chance the next prime $p_{n+1}$ will end in 9. Isn't that weird? It's pretty weird. We expect most things dealing with primes to be inherently random.

The paper puts together a heuristic explaining this distribution, but I'm not exceptionally interested in that- I'm interested in visualizing the bias. If you would like to read more about the discovery and back story, you can do so here. By setting the last digit of $p_n$ against the last digit of $p_{n+1}$ we can plot this phenomenon. By using a darker color for the frequency of that grid position when examining primes, we get a nice visualization. In addition, we can change the system of counting or 'base' to produce even nicer patterns. So without further ado here's an interactive of consecutive prime distribution:

Neat huh? Really nifty patterns. I ran the Sieve of Eratosthenes over integers up to 100,000,000 yielding 5,761,455 primes. Here's that terrible, terrible code. I decided to take this further and checkout the distribution of 3 consecutive primes. Any longer sequences become dramatically more difficult to visualize. Here is said interactive distribution (click and drag and whatnot):

On mobile or have a limited screen? Checkout the raw visualization for these guys: here for dim=2 and here for dim=3

I'd be interested to explore these distributions over non-traditional bases. For instance, consider a system $n_{base 10} = d_4\cdot 21 + d_3\cdot 10 + d_2\cdot 3 + d_1\cdot 2 + d_0$ for each digit d of a number $n_{base\ strange}$, or something crazy like that. I'll update if I have time to explore this.