Temperature histogram

In commemoration of the coldest dive I have ever done in a wetsuit, this is a histogram of the water temperatures I have encountered throughout my 160+ dives. Red is for warm water dives (i.e. in a 3mm wetsuit or shorty) and blue is for cold water dives (7mm suit and hood).

At 5ºC, yesterday's dive was a full 3ºC below my previous low (~8ºC at Friday Harbor).

Graphs of a blog of graphs

It's about time this happened, because there was no way I could let these pretty graphs just sit there. Blogger records visitor stats for the blogs it hosts and graphs them. So there are graphs of this blog's visitor traffic. Fig. 1 is a graph of visitor operating systems.

Fig. 1. Visitor operating systems for this blog

Compare that with the OS composition of another more high-traffic Blogger-hosted blog with stats I can access: the TA-run Brown ecology class blog (Fig. 2). Brown is rather maccy in general, but this is a really big difference!!

Fig. 2. Visitor operating systems for Brown ecology class blog

Finally, a general comparison is in order. As I did before when comparing my class diversity over time, this is basically a H' index. I wonder what Shannon et al. would think of my using their index to quantify diversity in ridiculous things, but it's the most familiar measure of diversity to me and here it is (Fig. 3).

Fig. 3. Diversity of this blog vs. the ecology blog

Of course, my real point in posting all these graphs is this idea: if I blog the graphs of the blog and then graph the blog of the graphs of the blog etc. etc. IT WOULD GO ON FOREVER!!!

Next time: web browser diversity!


As a side note: There is obviously some non-independence going on here (e.g. repeat visitors, of which I am sure there are many many many). But this is Blogger's fault and not mine and I am not making any statistical inferences here, so I'm going to pretend all is fine.

Operation retrieve dive gear

Success. Did not get lost or stuck waiting for the next 2.5-hourly bus!

Not the most efficient, but I am still really thankful for public transport. And now my gear is ready to help me freeze while running transects in Nahant this week.

Also: I never knew Providence had a 'Nahant Street' but apparently it does. It is a tiny little street off Charles Street. I passed it on the bus today.

Why "earth hour" annoys me

Because I see no reason to sit in the dark for an hour (other than making yourself feel all fuzzy inside) when there are a million other things you could do to cut energy use more effectively (and without cutting productivity). Like turning off the lights when you're not using them, and not running the hot water when you're not even touching the water, and not taking hour-long hot showers (I am especially bothered by people wasting water because it's like a 2 in 1 'water and energy' going down the drain).


* My room is lit by 2 x 36W fluorescent bulbs in a standing lamp.
** Taking an average shower flow rate of 2.5 gal/min, numbers from here. Note that this is calculated assuming 100% efficiency of the water heater, which is an impossibility, so any realistic value would be higher than what is graphed.

Off meal plan

This is a graph idea that came out of the inevitably frequent conversations with people on my hall and other 'off meal plan' upperclassmen about how ridiculously expensive the college meal plan is. I think it comes out to something like $9 a meal, but a number like that is not that easy to compare to the less well-defined manage-your-own-food-budget. I generally think of meal plan as 'a lot' more expensive than off meal plan but have no clear conception of what 'a lot' is. I decided it would be easier to visualise if I graphed it.

I was pretty surprised at how much 'a lot' really was. My initial idea was to stack things up on my 'off meal plan' bar until it equaled the meal plan bar but I pretty much ran out of 'large expense' things to stack, with over $200 to go.

The graph isn't meant to convey an expenditure switch because none of the other expenses actually 'came out' of meal plan money - Benthics $$ came out of a Brown travel grant and TA stipend, NEAS from the stipend and a research assistant job, dive gear came out of a summer research fellowship, etc. - it's just a visual comparison of magnitude.

Other (qualitative) benefits of not paying the college to make food for you:
1. flexibility
2. feeling like a real person
3. not being utterly lost and helpless when the dining halls close (spring break week, ha.)

Okay I am off to make lunch and work my butt off so I can spend the second half of next spring break freezing in Nahant water and looking for algae.

In and out

I've been scrambling to catch up with everything ever since I got back from the Benthic Ecology Meeting five days ago so I haven't had the time to mess around with data and graphs. But now it's spring break, and even though I have a to-do list as long as my arm, there's some time to play around with excel.

Like most conferences that aren't tiny, Benthics has concurrent sessions (4 of them) and they were close enough that it was very easy to switch from one session to another between talks. Each session had 5-10 talks, so 4-9 potential switches in total. Here is a graph of how much switching we all did (as a proportion of how many possible switches there were):

FYI, we are almost but not 'significantly different' from each other (p~0.09). Though I bet if I just ran a paired test between me and Jason we'd be different. I think I just want to believe that I am significantly different from smelly Jason :)

And below is another graph of switches (again, proportion of total possible), this time by session. Yes, there was one session in which I switched rooms at every possible time, though I think it was mostly moving back and forth between two of the rooms.


Also, 2 great graph ideas that may never see the light of ordinate and abscissa:

1. I wanted to graph our individual paths and movements between sessions, but we moved around and crossed between rooms so much that the graph is just a crazy mess of lines which fails to convey the point. Even the above graph (switching by section) is a little but crazy and falls on the very limit of what I consider worth graphing.

2. If I could figure out which combination of talks each person at Benthics attended, I could use some kind of ordination technique (I'm not sure exactly which would be most suitable) to characterise each person relative to everyone else based on that. Then people who cluster together would probably have similar interests and should talk to each other => awesome way to network, right? But it's kind of lame to do this when I only have data for 4 people. Maybe if I go to Benthics again next year I will try and do this...

Networking at Benthics

(estimated in approximate number of hours spent)

I think we are all terrible at networking.