converged.yt

Year notes: 2017

Intro

I’m back with a yearnotes for 2017. There were no 2016 notes, sorry! Hopefully these will make up for their absence…

What did I work on (for money)?

This year I quit my job working for NOAA in Woods Hole after some deliberation. It wasn’t a good fit for what I wanted to be doing. So, in May I stopped thinking about right whales. I then went to Hobart, Tasmania to work for CSIRO on close-kin mark-recapture. I was in Australia for about 5 weeks, which was great. From August through December I was working in St Andrews but just as an honorary research fellow. That was still great but I’m ready to have a proper job again. (And I do, see below.)

Papers

It was not the best year for papers… Things crept along with our big variance propagation paper, which we’re almost ready to submit now. Otherwise, our paper on the Distance package that I wrote got accepted (finally!):

The chapter I helped Steve and Eric with on distance sampling for Lenny Brennan’s book went to the publishers:

and Mark and I wrote a little thing for the Whaling Commission about when to do spatial modelling:

Coming soon (hopefully), Julie asked me to help out on an analysis for a fish paper:

and Roland and co terrified me with their efficiency in writing this review of hidden Markov model stuff:

I still have a lot of half-finished stuff, but it’s looking more half-finished (less half-finished?) than it was at the start of the year.

I didn’t finish writing my book, but it’s getting there…

What did I work on (when I could have been working for money)?

I wrote some silly Twitter bots this year… The below is adapted from an interview I gave about this stuff…

I also got interviewed by usesthis which was super-fun!

@birdcolourbot

@birdcolourbot was by far the most successful. It even got some coverage in the New York Times. I’ve since got quite a bit of attention for it.

The idea for this came from thinking about how Cornell Ornithology Lab’s Merlin bird identification app “thinks” about birds. The app takes your location and a description of the bird’s colour/size/activity and tries to tell you what you’re looking at. For it, birds are just probabilities of given colours occurring. I’m red/green colourblind, so I’m interested in colour perception and how different people see birds (or anything really).

I extracted plumage colour data from the Merlin app and used this to build blocks of colour (with width in proportion to the probabilities of a given bird being that colour). The only random element of this bot is which bird is posted each time. The bot as caused some consternation in the ornithology community on twitter, some folks pointing out that some of the colours are “not right” but this is partly because the data includes plumages for juvenile birds as well as adults.

(Sorry for borrowing your data Cornell <3)

@transect575

@transect575 generates haiku’s from whale surveys…

I’ve spent a lot of my time analysing the huge data sets collected by institutions like NOAA. Looking at these data one day, my computer returned the error:

[687, 23]: expecting numeric: got 'SURFACED IN BAD WEATHER, LOST AT SEA'

I thought this was somewhat poetic and that got me thinking, what if I put together poetry from these comments? I took data from surveys looking for the highly endangered north Atlantic right whale. When the observers make observations they also record notes in their data, things like weather condition, whether the whale is entangled in gear (a serious concern) or details of other non-whale phenomena, like fishing boats, weather etc. Given Twitter’s limit on length, the haiku (5 syllables, 7 syllables, 5 syllables) seemed like a logical poetry format. So I went through the comments, first anonymising them and removing duplicate or boring entries, then calculating the length (in syllables) of each comment. I then used cheapbotsdonequick to randomly pick comments of the right length and post them every few hours.

A large number of the haiku include information about entangled whales — a major issue in the Gulf of Maine and an important cause of mortality in right whales. I hope the bot reminds folks about the continued impact of humans on the ocean.

@clapping_bot

@clapping_bot tweets visual representations of excerpts from Steve Reich’s Clapping music

Where did I go?

Using a little script here and making a list of where I went, I made a wee map.

map of the world with 2017 destinations highlighted

This was a lot of travel. Hopefully I can do less in 2018…

I also went back to using an iPhone this year, which meant that I was able to do a little analysis of the step data. Some plots:

whole year of steps steps aggregated at the day level steps at the month level

So, it looks like I’m walking more, but I could still be doing better. My back is definitely feeling the benefit (he says, writing this from the back seat of a car…)

(Note that I’m not 100% sure on the day boundaries here, Apple appears to record everything in UTC, but I don’t see how to work out what timezone I was in at a given time (except by manually working that out), so that kind of sucks.)

Birbs

I really got into birding this year. I set a modest(?) goal to see 200 species during the year, and I managed 207 (though I did cheat somewhat by visiting 3 continents, see below). I use eBird to track what I’ve seen and using the data export feature, I put together a little script to build a species accumulation curve.

personal species accumulation curve for 2017

here the black line is my overall number of species seen, blue is for 2017 alone. Colour coding is for location (where AU is Australia, IC is Iceland, NO is Norway, NS is Nov Scotia, SL/IT is Slovenia and Italy and UK is… UK, the rest are US states.)

Who do I know IRL and on Twitter?

Back in 2015 I made a little table showing all the people I knew both on Twitter and in real life. I made this again below but I’m realising that Twitter is becoming such a mainstream thing that perhaps this is less interesting than it once was… (in particular I’m now meeting more people IRL who have Twitter than people from Twitter who I then meet up with in real life.) In any case the people who I do meet from Twitter remain charming and excellent folks. Here you all are:

(Code to generate this is in this gist.)

(Did I miss you?! I’m very sorry! Toot me and I’ll fix this!)

What’s next?

As of January 3rd I start working at CREEM again at the University of St Andrews. I’m spending the next 4 years trying to make spatial modelling better for folks who want to estimate animal abundance (in particular whales and dolphins).

That’s all for now…

Other posts

Year notes: 2017

Year notes: 2015

Reflections on teaching "Spatial models for distance sampling data"