Guest post: Personal journeys towards developing quantitative skills

This post is by Isla Myers-Smith and Gergana Daskalova, from the University of Edinburgh

******

Gaining quantitative skills takes you on a journey. When we start, many of us feel like we are behind and can never catch up. Those who feel too overwhelmed may never start the journey at all. And if we want to enhance diversity within the field of quantitative ecology, we need to overcome the fear factor in quantitative training. Reflections on our own quantitative journeys highlight that the major roadblock is taking that first step to bridge the quantitative skills gap. In the following blog post, we tell two interwoven stories of personal journeys towards developing quantitative skills to highlight how things can be different for the next generations of ecologists.

The quantitative skills gap

The mid-career academic: Half way through my PhD, I did not know how to program or to carry out my statistical analyses. I watched the students around me talk about the statistical programming language R and the analyses they were conducting. I was still working away in Excel. I felt so far behind that I thought there was no way I could catch up. I thought that my quantitative skills gap might be the biggest roadblock to my potential future career as an ecologist. One day I turned to a friend to ask for help. And she spent one hour with me teaching the basics of R. Across the rest of my PhD and postdoc, I slowly began to realize that coding and statistics weren’t that hard and maybe quantitative skills weren’t that scary after all.

The current PhD student: During my undergraduate exchange year in Australia, I got to experience so many new things – landscapes, wildlife, and cultures all coming together in an exciting and sometimes scary mix. And while there was no language barrier for me in day-to-day life, as soon as our ecology practicals kicked in, there was another language to master – R! All of the other undergrads around me knew what R was and I didn’t. We hadn’t studied any R in my first two years in Edinburgh, the whole thing felt very foreign to me and I felt very lost. With the help of my fellow students, I got the answers to the report questions, but I didn’t really know what was going on and I tried to put R behind me. Once back to Edinburgh and in my final year of my undergrad, a lecturer asked the class if anyone had heard of “R”. Nobody said anything. Eventually I raised my hand, since I had indeed heard of R. Suddenly, I became the go to person to ask R programming questions. I better learn this, I thought.

Everyone has their own journey to learning quantitative skills. As a mid-career academic and current PhD student our goal has been to try to make that journey a bit smoother for the students that we teach – to bridge the quantitative skills gap. We try to encourage student’s confidence that their perceived skills are an accurate representation of their actual skills.

Becoming a quantitative ecologist

The mid-career academic: Over a decade after my PhD quantitative catch up, I would now call myself a quantitative ‘big data’ ecologist. I love the way that code can let me conduct analyses on scales that otherwise would not be possible – to understand biome- and global-scale patterns and processes. The joy of compiling data, designing statistical models and running code and seeing an answer to a scientific question is perhaps my favorite part of science. I gain huge satisfaction from being out in the field collecting new data points all the way up to compiling cross-site datasets that allow us to understand how for example the tundra biome is responding to global change. Quantitative skills are a critical part of my science and I don’t know where I would be if I hadn’t learned them.

The current PhD student: I jumped into my PhD ready for action. For most of my PhD research I am working with large compilations of ecological time series, which meant that I could start testing my predictions early on. If anything, I had to force myself to not go ahead and just code, but to actually carve out time to think through my research questions and the theory behind them. The actual analysis part felt like an oddly comfortable challenge – oh, there were and still are many lines of code where I’ve stumbled, but for the most part, I felt okay about being able to eventually find the solution to my coding challenges. Now in the third year of my PhD, I have thousands of lines of code behind me. I have organizing many coding workshops and been teaching others quantitative skills for a few years. But, I still get a thrill every time I learn something new myself, or teach others and see the spark in their eyes when they start believing in their own quantitative abilities.

Quantitative setbacks along the way

The mid-career academic: I wasn’t totally new to the idea of coding when I started my PhD, I had taken a course in C++ during my MSc degree. The first half of the course was taught by a computer programmer working in industry. He designed group activities where we built simple games and had to pitch our games to our classmates. He gave us insight into how the tech industry worked. The course was fun! We were all doing really well. In the second half of the course, we were taught by an academic computer scientist. The course became about the theory of programming and I started to do really badly – I was close to failing that second half of the course and only just scraped through. I thought then that quantitative skills weren’t for me. I set out to avoid them in my science.

The current PhD student: I went to a math-focused middle school, for which I had to sit an entry exam. I felt like many of my new classmates were much smarter than I was. My first week also brought my first low mark which shook my world a bit. At my new school I could learn computer programming in the language Pascal. At the time I had only ever used Paint on my cousin’s computer once or twice and I didn’t really know how to use computers at all. A couple months into my programming classes, there was a competition – I made it to the nationals, and I got 4^th place in the country. I was also the 1^st placed girl. More competitions followed – we got to travel to other cities, and I was on the school’s programming team. We wrote our programs on floppy disks! I am not that old, but our school equipment sure was. I didn’t really think much about the fact that the competitions were dominated by boys. Initially I was thrilled to complete, but after a run of not ranking near the top, I dropped out of the team and focused on French and literature in high school.

Overcoming code fear and statistics anxiety

The mid-career academic: At the start of my quantitative journey, I was experiencing what we might call code fear and statistics anxiety. I knew from my high school and university courses that I had no natural ability in math or computer programming. Clearly quantitative skills were not for me. I looked around me and my instructors and fellow students who were good at quantitative skills seemed different from me – they were often male. I didn’t see myself in that part of my discipline. But, I came from a background of having two academic parents and being around science my entire life. How does it feel when you come from a less academic background? When you don’t see anyone like you as an academic mentor? How scary does acquiring quantitative skills feel then?

The current PhD student: My quantitative journey in my undergrad was quite a steep one – my final year was packed with lots of code trials and eventually successes. Code sort of felt familiar – I had spent my high school years studying French and Russian, so I knew how to learn a language and R was after all just another language. Months later, I remembered that I had actually learned coding 10 years before in middle school! The whole realization came to me as a surprise. I didn’t imagine that my middle-school coding would come in handy in my university ecology program. There was a general sense of frustration about quantitative skills among the students in my program. We made it through our final year and handed in our dissertations. On our final day, we were talking about how great it would have been to have had a friendly place where we could learn quantitative skills without the pressure of assessment. This conversation was the beginning of Coding Club.

Rethinking quantitative training

The mid-career academic: When I started lecturing, I was given non-quantitative teaching. I liked the courses I was teaching about conservation science and outreach and engagement, but I was wondering if there was teaching that I could do that would have a bigger impact. I thought about my own academic trajectory. What if I had been introduced to quantitative skills when I was in my undergrad? What if I had started my PhD already having those skills in place? What if I never faced the quantitative skills gap? Where would I be now? How would my confidence in my own science and career have been different?

The current PhD student: I find quantitative skills incredibly empowering. I particularly like how they combine two seemingly contrasting feelings – independence and collaboration. Being able to lead a research team throughout the whole scientific process, to swiftly do data tasks that before baffled you for days and to see your figures and findings appear on the screen inspires me even in the gloomiest of Scottish winter days and gives me the confidence that even if I don’t know the answer now, I’ll eventually figure it out. And alongside the independence, there is so much collaboration – people sharing tips, answering questions – online and in person. Being a quantitative ecologist is like joining in a wider field of people aiming to harness the power of data and communicate interesting findings about our changing world, regardless of the specific discipline. I like feeling like I am a part of something larger than just me and my own research.

Coding Club audiences are growing around the world. In this map you can see the users per city, with the larger darker blue circles indicating cities where we are approaching 10,000 users. Our aim by providing free quantitative training is to reach diverse user groups and overcome barriers to quantitative skills development.

It was these parallel quantitative journeys that led us to launch Coding Club. A peer-to-peer teaching initiative that was outside of assessed courses. Where people from any background, experience level and career stage could learn together. Where the learners become the teachers. When we started Coding Club, we didn’t know whether it would fill a niche. All we were building from was our own experiences in gaining quantitative skills.

Five years later – Coding Club has been a big success. The website has reached over half a million page views from over 300,000 users. We have started a new course at the University of Edinburgh and in the past couple of weeks we have launched a new free online course for global audiences. Our goal is to provide that easy start to people beginning their quantitative journey or those wanting to upgrade their skills. We want to overcome “code fear” and “statistics anxiety” to help more diverse people access quantitative skills around the world.

Our FREE online #DataScience @our_codingclub course is live! Join #EcoDataScience and learn about #Rstats, #dataviz, #tidyverse and more. Our self-paced course is for learners from diverse backgrounds. @DataLabScotland @EdinburghUnihttps://t.co/jOCb53WurN pic.twitter.com/kRpWS7Ua7q

— Coding Club (@our_codingclub) February 24, 2020

Here’s what we have learned so far from our experiences teaching quantitative skills:

1. Try not to call quantitative skills ‘hard’. Coding is something that anyone can master. We can be teaching coding skills to 10-year-olds. Statistics takes some thinking, but there is nothing fundamentally hard about say Bayesian hierarchical modelling. It all just depends on your perspective. When you call something ‘hard’, people disengage. When we acknowledge that it takes time, but we can all learn quantitative skills, you break down barriers.
2. Teach quantitative skills with a scientific question. Quantitative skills are just tools. Teaching a tool in isolation from how that tool is used to answer a real-word question is way less engaging. A data present is a great reward for learning a quantitative skill! Live coding and passive teaching approaches have much lower retention. Dropping people into the deep end and getting actively involved in problem solving is how you teach people to learn for themselves.
3. Start by teaching open science best practice. Though perhaps a more intimidating place to start, we can begin with teaching GitHub for version control, study pre-registration, and critical thinking about p-hacking and the reproducibility crisis. We can encourage sharing of code and data and a more collaborative approach to science. We can begin the training of the next generation of quantitative scientists by teaching open science best practice from the start!

Our Data Science for Ecologists and Environmental Scientists course is designed to be free and self-paced. We teach key skill sets across three streams – ‘Stats from Scratch’, ‘Wiz of Data Vis’ and ‘Mastering Modelling’ and 16 two-hour tutorials including data manipulation, data visualization and statistics, but also open science best practice. We have set three challenges using real-world data for learners to test their data science skills. The ecology-focused research questions are: Where are the red squirrels?, Where are Scotland’s high-value conservation habitats? And how does climate influence seabird breeding?

And here is what we have seen in the Ecological and Environmental Sciences program at the University of Edinburgh:

A wide variety of students representing all different types of diverse backgrounds are now attending Coding Club, taking quantitative courses, conducting quantitative research as a part of their dissertations, and thinking about or launching careers in data science. We see participation from more women, mature students and students from diverse backgrounds. Our sample sizes are small, but anecdotally, quantitative training has been transformational in reshaping student trajectories.

The biggest transformations that we see in our own students is the increase in their personal confidence. Along with learning quantitative skills comes the confidence in one’s own abilities to learn any new skill. Coding Club graduates have gone on to diverse careers including in the field of data science. They have learned to pitch themselves as lifelong learners who aren’t afraid of any particular barrier or anything that seems ‘hard’. Our students are taking their quantitative skills farther than we could have ever imagined at the start!

Check out our new online Data Science for Ecologists and Environmental Scientists course and let us know what you think in the comments below:

• What is the story of your own quantitative journey? What were the roadblocks for you at the start?
• If you haven’t started your quantitative journey yet or if you feel like you are behind, what do you feel is stopping you from progressing?
• How do you think we can best teach quantitative skills? Do you have any success stories or resources to share?
• How do you think we can increase diversity in quantitative ecology? How can we make sure that training in quantitative skills are accessible for everyone?

And if you have any further thoughts or you want to contribute please get in touch!