Great stuff, but please normalise to UTC in future. Everyone (reading this) knows their own offset from UTC and can probably makes sense of the data directly in that format. For anything else I have to go and lookup the offset between that timezone and mine.
For a website with a US centric audience, I think it makes perfect sense to use a US timezone: The graphs match up well with the sleep/wake patterns in the EST timezone. This makes interpreting the data a bit easier, i.e you can easily tell that there is a morning rush, a lunch rush, etc. (although interestingly, the HN audience doesn't seem to have well-defined break hours as there is just one large peak per day).
Not everyone knows the difference between their timezone and UTC. I always end up looking up the difference, and it really doesn't matter to me what timezone people use, so long as they state it.
