My university, King’s College London, offers a special Spatial Data Science (SDS) pathway with their geography degree that is open to both human and physical geographers. I value SDS and am taking two modules involving it, but ended up not choosing the pathway. However, SDS is a great skill to have, so the following interviews are a collaboration with my SDS peers, @nsu and @wootubejr, are to inform anybody who is interested in the pathway!
I am joined today by my friends Kai and Nazifa who are both third year BSc Geography and Environmental Science students at KCL like myself. What made you two interested in the SDS pathway?
Kai: The course provided a lot of unique support for teaching python and outlined a rigour for mathematics and statistics. I saw it both as a good challenge and a chance to learn coding.
Nazifa: I’ve always been interested in data and statistics, and the module offered the opportunity to look at data from a spatial perspective. I also had very little coding experience, so it was a chance for me to learn something new!
If someone is on the fence about this pathway, what sort of skills or mindset do you recommend for getting started in SDS?
Nazifa: A computer will only do what you tell it to do. This can be very hard for people who think sporadically (like me), but if you go into class with this mindset, things that seem difficult to achieve can be done by going step by step. This is why I enjoy SDS because it’s like problem solving in a way- it’s challenging yes, but once it clicks, it clicks!
Kai: You’ll want to find joy in learning by experimenting and troubleshooting. There’s a lot of tech lingo to learn and a lot of running code you expect to work only for big red error boxes to appear. Understanding these errors is where you make the most learning and luckily there’s lots of online support (e.g. Stackoverflow), but if wrapping your brain around input errors and data type errors is not for you it won’t be fun. That being said, when you get it right, it is very satisfying and feels like you made something unique with your own hands and mind.
It seems like you need some patience and flexibility for these courses! Now that you guys have that mentality, what are some examples of work that you have achieved in your SDS classes?
Kai: A lot of my coursework has focused on finding spatial patterns either in locations or specific variables. Last year, I identified the hotspots in Hong Kong for best and worst waste management, and this year I found that energy consumption behaviour in London households is significantly affected by the location of the home and can be proxied through citizens’ quality of employment.
Nazifa: For my coursework last year, I was able to look at butterfly geographies in the UK. By using something called spatial autocorrelation, I was able to find ecological hotspots of butterfly diversity in the UK and how this compared to hotspots of vulnerable butterfly species in the UK. The results showed that areas with high butterfly richness didn’t always correlate to where the vulnerable species were. So, if conservation efforts were to only focus on areas of high biodiversity, we risk leaving vulnerable species behind.
Those are some really interesting geographical insights! As previously mentioned, I’ve done SDS work, but am not on this pathway, so how do you feel this work differs from other GIS and data science work you have done in the past?
Nazifa: SDS is a bit different to remote sensing and other GIS work. There are some skills that will definitely overlap. In that case it becomes a difference between software and manually clicking buttons as opposed to writing code. But, as it is in the name, a big focus of spatial data is the data itself: how we as data science chose to analyze it, represent it, and what it means. For example, for one of the very first courseworks, we had to make two maps representing the same data, but telling different stories; what classification, colors, etc. were all important in changing how people interpret your map. While this is talked about in other classes, it is more stressed in this module.
Kai: Through SDS you learn a lot about making your own data frames and applying unique spatial methodologies which other modules won’t teach. Most GIS work goes into land-classification and aggregating statistics spatially, but SDS teaches you how to make fair statistical representations and spatially dependent models. Many statistics will be extensive, meaning the bigger or more populated the place is the higher it will be, and many variables are affected by the neighbours around it which is sensitive to the scale you measure (think gerrymandering). SDS teaches you to factor these biases through methods like spatial autocorrelation and geographically weighted analysis, bridging the ‘spatial’ and ‘data science’ in SDS.
Also on the topic of comparisons, do you find this subject more difficult than others or other coursework you’ve had to complete?
Kai: Yes and no. On the one hand, you need to complete every practical assignment to succeed and conceptually I think it’s one of the least straightforward. The reward however is that once you learn the methods, a lot of the courseworks relies on the application of these methods and thus don’t require extensive reading. It’s almost like doing the course week to week is harder, but the final assignments take less time and effort as a result.
Nazifa: Yes and it’s not a competition. You need to put a lot of extra work outside of teaching hours to make sure you get it. Luckily for me KCL’s SDS department is very good and there’s lots of help offered either through office hours or resources online.
It’s great to see that you guys have worked through these difficulties in your coursework, yet it’s not just academically that SDS comes in hand. How are you two planning to use these skills in the future? Is it useful for job applications or further education?
Kai: SDS has helped give me a clear vision for my future career paths. I applied to an Imperial College London data science course because I felt SDS prepared me well despite my geography background. Additionally, a lot of environmental science consulting and R&D is based on analytics. Even just generally, I find myself able to make better connections because, even if I come from geography, everybody is talking about ‘quant’ which I am able to keep up with.
Nazifa: Yes, I found it has been useful for me! I was able to transfer a lot of skills I learned in SDS to the research fellowship I did over the summer where I had to work through a very large data set.
Thank you so much for joining me! Do you have any final tips for future geographers?
Nazifa: Geography is a bit about everything, so make sure you stay informed and read current news. The more you think about geography and are able to make those everyday connections in your real life, the easier the subject becomes (in my opinion at least, but I’m also just a nerd).
Kai: Geography is in a great place right now because so many pressing issues seek knowledge of spatial-scales and the environment. Stay knowledgeable and well informed on global issues, knowledge gaps and geographic applications and people will respect you even if you’re not a doctor, engineer or lawyer.