The second month of Gillian Rowan’s European Space Agency (ESA) internship brought with it breakthroughs, breakdowns (of code and hair symmetry), and some breathtaking weekends away. From the rocky beaches of Polignano to the marble halls of Florence, here’s how she learned that stepping away from the screen is often the key to coming back stronger.
Month 2: Finding my footing in Frascati
This month was AMAZING! Having spent my first few weekends in Rome, I decided that it’s about time to venture a little further out and see some of the rest of this beautiful country. Join me for a glimpse into my time exploring the stunning coastlines of Puglia, taking in the artistic treasures of Florence, and of course sitting at my laptop trying to figure out what on Earth all of these error messages mean!
Week 5: Puglia Dreams and Research Reality
Everyone keeps talking about how beautiful the southeast coast is – so I caught a train right after work on Friday to Bari. And wow! What a decision that turned out to be! The old town is everything you’d picture in a Mediterranean dream: light stone buildings and streets, flower gardens, and women lining the alleys outside their homes hand-making orecchiette to sell. Since it’s one of my favourite dishes, I went to the restaurant where Spaghetti all’Assassina was invented for dinner. Needless to say: their version put mine to shame! I have some recipe tweaking to do when I get home to Brisbane. The next day I drove down the coast and finally understood what all the fuss was about. Polignano a Mare was breathtaking. Just north of town, I found a “beach” where the shore was sunken ruins instead of sand. Only in Italy, right?
Back at the office, I got back to work on making my samples of radar images to train our AI model on. Clearing my head over the weekend was a good choice: things just clicked this week. I got two big chunks of my code working and produced the first of my radar image samples! I still have to work out how to handle the images’ metadata and how to automate the process but finally getting my code working and seeing the first data samples has me on cloud nine!
Week 6: Making progress
Still catching up on my lack of sleep from last weekend, I decided to stay near Rome this week. I brought my bike into the city with me and set out to explore some of the neighbourhoods outside the city centre. It was incredible to see so much Roman history casually integrated into daily life and urban spaces. On Sunday though, I made a big mistake: I got a haircut. The salon had great reviews, but I should have known when I walked in and only saw big curls that my pin-straight hair might be an issue. I think I’ll wear my hair up for a while, that way no one will see that the left side is a full two inches longer than the right.
The work week started wonderfully with two Zoom calls back to Australia – first celebrating a friend’s thesis submission, then catching up with my supervisor, Stuart. Getting his feedback on my work was really encouraging; he found a couple spots needing more explanation but was very positive overall. Hopefully that paper can be my first PhD submission soon! I also met with my ϕ-lab mentor team this week about my progress and processing decisions. They gave me a few considerations to keep in mind while reading the literature and digging into the models. The most important and challenging one was evaluation: how will we determine if the model is producing accurate and consistent results? With a radar image that has multiple types of information, it’s not as easy as just checking if the pictures look alike. I think it’s back to the books for me.
Week 7: I now understand people’s love of GitHub
With a long weekend coming up, I decided to go to Florence. I kept catching glimpses of the famous Duomo between buildings and partook in the classic Florentine tradition of “wine windows” and had a couple glasses of wine on the streets with friendly strangers. It felt like summer in Montreal and made me a bit homesick. On Sunday morning, I woke up strategically early (which in Italy is a breezy 7:15 am) to be among the first into the Galleria Uffizi. Wandering through the museum with no crowds was great – I had works by Bernini, Caravaggio, even Da Vinci, all to myself. 20 km of walking around the city finished with a free meal at the central food market. From what I could decipher, I think they mistook me for a Macedonian travel blogger…
You know those days when your brain just doesn’t work? That was Tuesday and Wednesday. I was adjusting my scripts to store information from each of the processing steps in a single file. And since it takes information from each processing step, I was making changes to each processing script. Now, if you’ve ever done any coding, I’m sure you can guess where this is going. But being the somewhat inexperienced coder that I am, I made all the changes I thought were needed and clicked “Run” … Everything broke. All of it. Considering how many changes I had made I also had no idea where to start looking for the issue. So, instead of trying to parse through everything, I reverted everything to the previous (functioning) version of my scripts and started over, piece by piece. And that, my friends, is how I came to love GitHub.
Week 8: It’s all coming together
This week was all about making my pipeline fit together nicely. I finished all the individual scripts, but now the outputs of one need to align with the inputs of the next. The variables need consistency throughout, and very importantly, everything needs to be tidy. Tidy code lets other people read and understand it more easily and reduces computational demand – which is crucial when you’re working with big, complex datasets like mine. It’s been a bit tedious to comb through everything, but hopefully it’ll make my scripts more usable for myself and others in the future.
We had a huge disruption this week in the form of a giant nest of flying ants that was found in the office. From one day to the next, thousands of ants appeared. So, while exterminators were called in to take care of it, we took over a conference room. It felt like the office equivalent of a slumber party: not quite enough space, way too much stuff, and snacks everywhere. It was one of my favourite days so far in the ϕ-lab.
This was a big month for me. Though I’m sad to think that I’m running out of time at ESA, I’m so proud to have made such substantial progress this month. Getting to learn from the ϕ-lab team has opened my eyes to so many new ideas and applications of both AI and Earth Observation. And of course, Italy keeps exceeding all my expectations! Stay tuned for my last update on my time here in a couple of weeks.