Long-term volunteer Antonio Campello

“It takes a lot of work to make a problem ‘data-scienceable’ and it takes even more work to make data usable.”

Pronouns He/him
Roles Senior Data Scientist at Digital Science; I am a long-term volunteer, and I think I have quite literally taken on all possible roles at DataKind UK (meetup organiser, committee member, volunteer, troubleshooter, data ambassador, event host...). Most notably, I led the Scoping & Impact Committee for two years.
Links https://acampello.github.io | https://www.linkedin.com/in/antonio-campello-ds/

How did you first get into data?

I come from a very academic background. Having done a PhD in mathematics, and worked as a postdoc for a couple of years on theoretical maths (with pen, paper, and whiteboard!), I ventured into a bootcamp called Science to Data Science seven years ago. After six weeks of intensive data science, I decided it was the right choice for me and never looked back.

What do you wish you’d known when you started to get into data?

Problem first or ‘begin with the end in mind’. This is advice I give to every junior data scientist, but it is also something we used to discuss quite extensively at DataKind UK’s Scoping Committee: What actual organisational problem is your project trying to solve? Spend most of your time understanding how your project might impact the organisation, who are your users or stakeholders, and how it’s going to be applied or communicated.

Once you identify the problem, you can try to find a suitable solution. If a simple solution exists, that’s best for everyone! I have seen complicated data science solutions suggested for ‘non-problems’ more often than not. I’ve certainly been guilty of this numerous times, especially at the beginning of my career.

What was your first volunteering experience with DataKind UK like?

I was a volunteer at a DataDive weekend in 2018 that looked into categorising charity sectors in 360Giving’s data resource. It was a packed day, and overwhelming at times. It went by quickly, and I felt the hackathon-like adrenaline of finishing a piece of code in a couple of hours!

The energy in the room was so positive, with like-minded volunteers and charity representatives working towards the same goal, that I knew I wanted to be involved with DataKind for the long run!

What are the most important things you learned from volunteering?

I learned a lot about what types of analyses are the most sought-after in the charity sector, as well as the various journeys towards data maturity in the sector. In 2020, we wrote a full report on the Data Maturity of DataKind UK’s partners. Sometimes what they need is one graph that will be the start of a compelling narrative, sometimes it’s some data cleaning, sometimes it’s something more complicated: there’s value in everything.

My biggest learning as a Data Ambassador and a Chapter Lead was how to bridge the gap between talented data scientists that want to use the latest technologies and passionate charity representatives. It takes a lot of work to make a problem ‘data-scienceable’ and it takes even more work to make data usable. Managing expectations and managing volunteers is definitely a skill I have transferred to my professional life.

What is a data project that inspires you?

The list is endless — every day someone comes up with a new idea for using data science for the public good, and I think that’s great! Among the DataKind projects I participated in, one of my favourites was a project with Citizens Advice Lewisham (CAL) to correlate deprivation with the services they provide. It ended up being presented to the Mayor of the borough, and was received very positively to help CAL progress their mission to serve people in need in Lewisham.

At the risk of ‘national’ bias (and because I cannot resist talking about another project!), one of my favourite non-DataKind UK projects is Operação Serenata de Amor, developed by a team of Brazilian data scientists to contest dodgy expense claims filed by politicians. It automatically scans receipts and claims, runs them through a database, and classifies them as suspicious or not. In addition to raising public awareness of how money is spent, more than BRL 3.6 (approx £500,000) has been contested as suspicious reimbursements since it began.

Is there a resource you’d recommend?

I occasionally listen to the Real Python Podcast. It makes for a good soundtrack for an hour-long exercise session.

Information is Beautiful (the books and the blogs) contains a plethora of accurate, aesthetically pleasing, and oftentimes mind-opening data visualisations.

Tell us something completely non-data-related about yourself!

I used to play piano at various train stations and airports as a hobby. Nowadays, I play to amuse the cat and our newborn baby.