It’s just over a year now since Ed Barter joined the Herdify team. We sat down with him and took a deep dive into his love of data.

Where did your career begin?

My career began at the University of Bristol. It’s where I did my PhD and got my first job as a Researcher. At some point in my time there, I went from being a student to having a career, but when exactly that was is hard to be precise about. 

What career moment are you most proud of?

In terms of standout moments, it would be having a paper on the cover of Royal Society Proceedings – the oldest scientific journal in the English speaking world.

The work was one of the first projects I’d done after getting a proper research job and it was great to be able to put some doubts about being able to do it at that level to bed.

And what challenges have you encountered?

In many ways, I’ve been very fortunate. In maths/data science, I’ve found something that not only do I enjoy but society values enough to make it a career. 

Working for start-ups has meant that my security is connected to the company’s fortunes. It is a challenge when things aren’t going well, but ultimately it’s worth it for the upsides of building something.

How would you describe data science to someone?

Data science is the process of finding stuff out from data. 

There are many different parts of that process, from getting the data to building models and implementing systems to deliver their results.

 There can be specialists who focus on each part of the process, but equally often, the same person is doing the whole thing. 

And how does that intersect with behavioural science?

Behavioural science tries to answer questions about why we, as people, do the things we do. Traditional behavioural science does this using carefully designed experiments that test specific scenarios. 

But we can also answer these questions about behaviour using data. Experiments involving large numbers of people can be expensive and difficult to control, so data science helps with questions about how people behave in groups.

What most excites you about your day-to-day job?

I like finding things out and solving problems and I’m lucky to do a job where I can spend a lot of my time doing this.

What advice would you give to someone wanting to start a career in data science?

The best advice is to practice data science in the wild. Choose some data to investigate and try to answer some questions you are interested in. 

There are lots of resources and tutorials available. They are great but tend to focus on the mechanics of implementation and not the research and discovery, which is what real data science is (or should be) about. 

So many data science projects fail because the problem to be solved gets forgotten early in the process, and the wrong tools are used. 

The most effective data scientists can take business problems, turn them into research questions and then use the data to answer them.

What motto do you live by – not necessarily just at work, but in life?

Don’t let the perfect be the enemy of the good. It sums up a lot of how I think about things, at work and in life. 

It speaks to the importance of keeping things in perspective – when something goes wrong and is not perfect, I like to ask “how can I make sure it’s still good?”. 

More specifically, at work, it helps ensure that the models we build have an impact. 

Data science is plagued by projects that fail to deliver. By asking “does this model help?” rather than “is this model perfect?” the right developments get implemented.

What do you see for the future of data science?

The last few years have been a boom time for data science, and it is experiencing some growing pains. Everyone has been wanting to do data science without asking what they need. 

In the future, we will see a distillation of components of data science and a better understanding that the vast majority of companies are a long way from needing the kind of deep learning capabilities Google has. 

We are beginning to see the end of the assumption that the key to data science is getting as much data as possible. 

Instead, people are focussing on getting the most out of their data; this is a trend I would love to see continue. 

Which living person do you most admire?

I admire people who can dedicate themselves unselfishly to helping others or society. That means I don’t know the names of the people I admire the most, as they do things to make the world a better place without promoting the fact they’re doing it.

Which historical figure do you most identify with?

Historical figures are far too notable to identify that strongly with, but I have always liked the idea of Galois – a mathematician interested in many things. Unfortunately for him, one of them was pistol shooting and he died in a duel aged 20.

Tell us about something that people might not attribute to data science.

A big part of data science is explaining the findings, often visualising data. 

An early pioneer of data visualisation was Florence Nightingale. Much of her success in gaining support for improvements in nursing is attributed to her ability to convey the impact on soldiers’ death rates. 

Nightingale’s most famous contribution, a kind of pie chart called a coxcomb plot, is used extensively to this day.

What’s next?

There’s a lot in the future for Herdify, all geared at helping our customers boost their marketing with the power of human influence and word-of-mouth. 

A few highlights are taking our technology international and helping brands that are launching and want to use word-of-mouth from day one.