From archaeology to data science: the joy of iterative career pathsDiscovering my love of all things dataLaura LewisBlockedUnblockFollowFollowingFeb 25At school I hadn’t planned on doing anything particularly technical as a career.
I took a maths A-level largely as a refreshing break from the essay-writing of history and English lit, and the time-consuming creativity of fine art.
I went to Cambridge to study Archaeology and Anthropology (another iterative decision process, but not one for now), and then on to Oxford for a Master’s and then a PhD in Archaeology.
Three examples of me as an archaeologist.
Left: Monte Polizzo, Sicily, 2006.
Centre: Atapuerca, Spain, 2010.
Right: Lakaton’i Anja, Madagascar, 2012 (right, foreground).
One of the great things about archaeology as an academic discipline is how broad a church it is.
You can study everything from the aesthetics of a Roman vase to radioisotope dating and still call yourself an archaeologist.
I found myself towards the more technical end of the spectrum (and very thankful for that maths A-level after all).
My field of interest was human evolution and the earliest dispersals of our species across the world.
I realised that a lot of our thinking on the topic was pretty qualitative — “this stone tool looks a bit like that other stone tool, so probably they’re the same people”.
Here, I decided, was a great opportunity to try using, you know, actual quantitative analysis, and maybe try and get some actual answers.
This is how I found myself in the incredibly niche field of archaeological statistics.
Not so much a field as a small patch of grubby ground containing me and maybe two other archaeology nerds.
But suddenly, by applying statistics to a brand new area, I was able to discover all kinds of important stuff about the past that we just couldn’t get at before.
My research contradicted some previously canon beliefs about our cognitive and cultural evolution, and about what happened when our species first spread out of Africa.
I was evangelical about the benefits of the rigorous use of statistics in archaeology.
I taught statistics to fellow archaeology students.
I spoke at conferences.
I published a book and a bunch of papers on the topic.
Trying to figure out how to have an impactful careerHowever, as much as I loved my subject area, I found myself frustrated about the lack of real world impact I was able to have.
One of the problems with niche fields of research (at least those without an obvious business or industrial application) is that, most of the time, only a small number of specialists actually benefit from your work.
I felt (and still feel) driven to work on something that actually benefits people in the real world, not in an academic ivory tower.
It’s that drive that made me decide to leave academia.
I loved the analytical part of my work, and I wanted to keep working with data, but I wanted to work somewhere where I could help solve real-world problems for actual people.
That’s why I next decided to join the UK Civil Service as a statistician, on the ‘fast stream’ (their accelerated development program for graduates) — I’d get to keep playing with data, but I’d get to use it to help make my country a better place.
My first post was in the Department of Health, working on financial and economics models for the NHS.
I loved getting to work on big juicy tangible problems, but quickly discovered the trade-off — working on nationally important problems means dealing with nation-sized bureaucracy, and a pace of work I found frustratingly slow.
I had managed to tick off my ‘data and analysis’ and ‘real-world impact’ boxes, but something else was missing.
Since moving to London, I had started to discover the wonderful world of startups, after going to meetups and meeting all kinds of amazing people building amazing new things.
I wanted to get closer to the action, so when a job came up at the Treasury to advise government on its venture capital policies, I applied straight away.
In this job I got to keep all the stuff I loved about data analysis, and keep making an impact on national policy, and now I got to add a genuine interest for the subject area.
I also got to apply statistics and analysis to a new area, being the first person in the 20-year history of my policy area (the Enterprise Investment Scheme (EIS) tax reliefs for startup investments, and related schemes and their acronyms (SEIS, VCTs, SITR and CITR) to conduct a statistically-rigorous evaluation of the schemes, part of which was published in the consultation on the government’s Patient Capital Review.
Finding the right working environmentAgain, I found much to enjoy in my job.
However, I still felt like the working culture wasn’t the right fit for me, and that I was still one step too far away from the action.
I decided to give working in a startup a try.
I joined a proptech startup, working my way up from pricing analyst, to head of revenue and analysis, to a sort of hybrid data product manager and head analyst.
I learnt so much here that it probably deserves its own blog post.
One of the things I’m proudest of is building up the analysis and reporting capacity of the company from scratch, moving from running off a confusing collection of spreadsheets and part of a database built for other purposes, to having a reporting infrastructure that serves everyone in the company and helps people make data-supported decisions every day.
I’d added a couple of new ticked boxes in my latest career iteration — working somewhere where you get to try out your new ideas all the time, and the kind of fast-paced collaborative culture that suits me.
However, I started to feel like I was still one small iteration away from where I really wanted to be.
Towards data scienceGetting to work with developers and delving deeper into the tech world, I began to discover that there was a whole extra world beyond data analysis, where you could use a combination of maths and coding to solve complex problems by building even more intelligent models.
For a little while I’d been feeling like there was a skill gap between where I was (e.
I designed a nearest-neighbour model for estimating nightly prices for Airbnb properties, but I couldn’t quite build one as sophisticated as the one I had in my head), and all the cool stuff I could see other people doing, which is where I wanted to be (e.
being able to design and build a model for predicting future Airbnb prices myself, testing it, and then working with developers to put it into production).
I found out that the term that encompassed the skills gap I had identified for myself was ‘Data Science’.
I started trying to teach myself data science in my spare time, but I in the end I decided to attend a data science bootcamp to get myself up to speed and start working with data science tools as quickly as possible.
A lot of people have asked me why I decided to go down the (more expensive) bootcamp route rather than teaching myself; the honest answer is excitement combined with impatience!.As of the time of writing, I am currently attending a full-time data science bootcamp at Flatiron School.
It’s a lot of hard work, but I’m learning a huge amount each day, and thoroughly enjoying it.
For me, I don’t see data science as one specific career decision.
Instead, I view my whole career path as a journey of many years towards data science, and I’ve learnt so much from each iterative step.
Not just technical skills and knowledge about different domains, but also about how I work and the values that are important to me in a career.
The thing about this iterative career journey is that each step gets smaller.
And I’m so excited to finish this last little step in data science up-skilling and start the rest of my future as a data scientist.
.. More details