Step 1: Apply A Reductionist Approach
A blog is made of posts so let's look at our history of posts. To make 1000 posts over 3 years, 9 months, and 7 days we averaged a post every 31.5 hours - not bad for a bunch of very busy grad students. If we look at the number of posts over time, we see that the distribution wasn't at all even.
Step 2: Look For Trends
If you take a squinty-eyed look at the history of our posts, you can make up a slight pattern. We seem to post more in the summer and less near the start of fall and the end of spring. If we plot the average number of posts per month over the calendar year, complete with a 1-sigma error region, we see that while there is something of a trend, it's hardly iron-clad.
Step 3: Derive A Metric
We know we're writing posts, but does anybody care? Luckily we can invent a quantity to measure how much discussion in generated by our posts, or the average number of comments per post.
Step 4: Fit To A Model
No piece of data analysis would be complete without some sort of attempt to fit our data to a model. What theoretical model best fits the growth of a blog made by a bunch of nerds from BYU? In a word, growth. Looking at the number of unique IP addresses visiting the blog per month since late 2008, we see nothing but roses.
To all the readers, thanks for putting up with the first 1000 posts. Here's to the next thousand.
Nick,
ReplyDeleteThanks for posting these statistics. These are really interesting.
And furthermore I would like to thank everyone involved in the blog whether writing posts, comments or just lurking around. To me you all make this a really enjoyable experience.
And now here's to making the next 1000 posts even better! Thanks again for everyone's support!
Congratulations! And may the next 1000 posts come in half the time until we're getting 1000 quality posts per day.
ReplyDeletebtw, based on these statistics, I have concluded that you are not using Google Analytics to track usage. Otherwise you could nail down a lot more specific information that might creep out your users.
How many unique users from Utah in the last three days
How many repeat users from Utah between 12:00 and 3:00 am in the last three days
How many repeat users from Utah that have left stupid, humorous or otherwise useless comments
How many repeat users from Provo visiting between the hours on 12:00 am and 3:00 am in the last three days while in their underwear leaving stupid, humorous or otherwise useless comments
etc...
Seriously, you can do that... well ok, minus the wardrobe selection. =:)
Love it! From the engineer's point of view, I propose we apply a Kalman filter to our model and account for noise as well.
ReplyDeleteSeriously though, I sincerely appreciate the opportunity to be a part of this blog. I really enjoy the posts and the authors here are top notch. Here's to the next 1000 and beyond!
I wanted to note again that this wasn't just me. We've been bouncing around some ideas about what to do for the 1000th post for a while. This was the result of a group effort.
ReplyDeletejmb275,
ReplyDeleteIt sounds like you've already started working on the follow-up paper that improves and/or completely disproves our model, which is of course step 5.
Congratulations to everyone for the first 1000 posts!
ReplyDeleteBy the way, Nick, you completely stole my idea for the 1000th post. You're just a whole lot faster at analyzing data than I am (no surprise there). I do have some questions, however. How often do we post on various topics (politics vs humor vs string theory, etc)? Which topics get the most comments? What correlations might be drawn between authors and topics they post about? How have things changed over time? What in the world happened in Aug 2009?
That is the wonderful thing about data, and the universe in general, no matter how much you look at it, there's always more to find out. Here's to another wonderful 1000 posts! Congratulations!
Bill,
ReplyDeleteThere is clearly a lot more analysis that could be done, but as it generally goes in science, being first is often more important than being right (or at least complete). One thing that really intrigued me is that the huge jump in comments per post in August 2009 wasn't accompanied by anything close to a corresponding jump in overall readers (or at least visitors). Apparently we just started talking more.
Congratulations on the good work. Keep it up!
ReplyDelete"a big jump starting August 2009."
Well, yeah. That's when I drew attention to you. :)
Actually, there is a lot of truth to what Jared* says. Yes we were "talking more", but as Jared*'s points out, he linked to our blog. And from that specific link (I remember) a bunch of others found us. Furthermore, through their comments here I "found" their blogs and left comments and... Traffic grows with extra links/comments to each others blogs.
ReplyDeleteYou know Jared, we may owe you a lot linking to that post as I believe it started a snowball effect. :)
Congrats!
ReplyDeleteI think it was Jared* that first sent me here. Good move, Jared*.