getting an OWID-like job

I often get messages from people (often students) who are interested in my work at Our World in Data (OWID) and want to know how they can get a job like this one. They (and maybe you) want to know how to prepare for such a role.

On this page I try to share some advice and some notes on the personal journey that brought me here. This is probably not exhaustive, but I reckoned I should write some words about this. I hope this is helpful!

what this work actually is

Working at OWID is incredibly rewarding. I am very privileged to be in this position, where I can invest my time and efforts to help make progress against the most challenging global issues. I also have amazing colleagues, who excel in their respective domains and from whom I learn every day.

Working at OWID can mean different things depending on the role. We are around 25-30 individuals, working on data collection, analysis, visualisations, communications, and various other tasks that contribute to our mission.

For me, I mostly work on importing and cleaning important datasets, maintain our data infrastructure in our etl, lead analytics to better understand our impact, etc. This means a lot of python, SQL, Claude, BigQuery, etc.

skills to develop to get hired

I’ve generally not been involved in hiring decisions at OWID, so I can’t provide a firsthand account of what actually gets someone hired. But I can share my perspective on this.

Different people at OWID have different sets of skills, but there are some common traits that we value a lot:

curiosity

At OWID there are lots of interesting discussions and articles being published. Being curious means actively engaging with these discussions, reading what our colleagues write, checking our comms, etc.

attention to detail

Being meticulous and thorough in your work, ensuring accuracy and reliability. Being wrong in a chart, indicator or article can have a huge impact on our reputation. So we really invest time and resources in making sure everything is perfect and error-free.

autonomous working

We are a remote team. This means that most of us are working autonomously on our own. We do collaborations, and team discussions, but most of the time we are expected to manage our own tasks and deliver results independently.

adaptability

In my personal case, I often need to distribute my time across different lanes of our work: data research, data engineering and analytics.

communication

Remote settings mean that synchronous time is often limited. Clear and effective communication is vital to keep team alignment.

how i got here

While doing my undergrad (Telecom Engineering), I was very interested in wireless communications and signal processing. During my masters (2015-2017), I focussed more on data processing and machine learning.

During my masters, I really wanted to go abroad. I looked into a program at my home university (UPC) that would allow me to spend a year at KTH in Stockholm. And I succeeded! During my stay in Sweden (2016-2017), I discovered Hans Rosling. He planted a seed in me, and fuelled my interest in using data visualisations to understand the world.

I was then lucky to get into a research program in Japan. There, I was able to contribute to open science, working on a custom CNN system (CNNs were big back then) to estimate typhoon intensity.

When I came back to Europe, I had lots of ideas on how my career could continue. Should I pursue further studies? Should I dive straight into the industry? An opportunity in my home town, Barcelona, appeared, and I decided to take it. I spent nearly 3 years working there as a data scientist. I learnt a great deal about practical data science and the daily challenges a company needs to face.

While I enjoyed my work in Barcelona, I started to feel that I wanted to work on projects that had a broader societal impact, rather than being confined to the commercial interests of a single company. It was COVID-19 times, and I was sending my CV daily to different orgs and companies I was interested in.

In my free time, though, I continued working on various open source projects. Contributing to the community was very important to me. Most of my studies had been subsidised by taxpayer money (like other fellow citizens in Europe) and/or scholarships, so I felt I had a responsibility to give back to others. I also genuinely enjoyed the process of collaborating with others.

In winter 2020, I started contributing to OWID’s Vaccination global dataset. They were compiling data from all countries in the world, and needed a script per country. I started contributing more and more. They decided to open a position to help them out properly, and I applied. The vacancy was posted as a GitHub issue! It was a fortunate turn of events. But I like to think that I had been working towards this without knowing.

other OWID-like orgs

There are a lot of resources out there to help you find and apply to roles at organisations similar to OWID.

I personally always recommend investing time in 80,000 Hours. They have extensive guides and podcasts on career planning and finding impactful work (they have a Job Board!).

Gapminder is also a non-profit organisation focused on promoting a fact-based worldview for everyone, with teaching materials, quizzes, etc.

Other orgs include Coefficient Giving, GiveWell, far.ai, GovAI, etc.

if you want to write to me

Yes, please do — I read everything, and I’d rather answer a question than not. A few things make it much more likely I can reply properly:

1. Ask one specific question. Something this page didn’t answer, about something we’ve published or something you’ve built. “What path would you recommend?” is a question a chatbot answers better than I can; “why does the series in X switch sources in 1990, and does that affect Y?” is one I can still do better.
2. Show me something you made. A notebook, a chart, a scraper, a blog post, a repo. It doesn’t have to be polished or finished. It tells me more in thirty seconds than a page of background does.
3. Project proposal. If you are working on a project where I could be useful, please pitch it to me, so I can hopefully help!