Total users over December 2015. I expected 70 users or so, maybe. By Day 2, almost 10,000 were signed up, which caused quite a bit of panic on the server capacity end of things.
The number of minutes it took the first 100 solvers each day to solve both parts of the puzzle. The three slowest days were Day 1 (red), which had a lot fewer users working on it at first than any other day, and Days 19 (green) and 22 (blue), which were two of the hardest puzzles.
The number of solvers each minute for the first 240 minutes after each puzzle unlocked. The days aren't labeled, because it's just a blob at this scale. Makes a nice curve, though; almost a Poisson distribution?
The number of cumulative solvers for the first 240 minutes after each puzzle unlocked. The two days with the most solvers in that timespan were Days 4 (green) and 10 (blue).
The number of users that solved both parts of a puzzle within 24 hours of the puzzle unlocking. Even though Day 1 was a "slow" day in the graphs above, by the time the whole 24 hours is finished, Day 1 still comes out as the most-solved-on-that-day puzzle.
The number of users with at least one wrong answer on a puzzle within 24 hours of the puzzle unlocking. Day 2 is the highest here, which I suspect is from the difficulty increasing to something that filtered out most of the non-programmers.
The number of users that solved both parts of a puzzle with no wrong answers within 24 hours of the puzzle unlocking. This isn't quite the difference of the above two graphs, as the above "wrong answers" graph is all users, whereas this one only subtracts users with wrong answers that also solved both parts. The number of perfect solves on Days 19 and 22 was very low.
