Project volunteer Melissa Torgbi

“I saw data science as a way to solve problems using technology… that’s what it gave me.”

Pronouns She/her

Roles Data Scientist at SAS
DataKind UK Data Ambassador on several DataDive projects

Links LinkedIn

Melissa volunteered with DataKind UK over the pandemic, and has since been a Data Ambassador on two charity projects. She tells us her top tips for future Data Ambassadors and about her route to data science through machine learning.

Why did you decide to become a Data Ambassador (DA)?

I found out about DataKind by searching for data for good projects online, because I wanted to use the skills that I have to help people. I volunteered for a virtual DataDive weekend and really enjoyed it. It was really nice to work with other people in groups to solve problems.

From there, I wanted to be more involved in creating the environment that I’d enjoyed at the weekend, and helping the charity (or social enterprise) behind the scenes. I wanted to get to a stage where you truly understand some of their challenges and problems, and learn how to create the questions and prepare the data in order to get weekend volunteers to provide real value and insights for the charity.

As a Data Ambassador, you’re one of the driving forces behind DataDive projects. The charity has some ideas and data, and you take them from ‘This is the data that you have, what does it need to look like for the weekend, what are the questions you want answered?’ to a weekend of analysis. When it comes to preparing for the role, every charity is different, so it would be hard to define exactly what each process could look like. You have different data, and the organisation has different goals. You have to see what they want to get from it, and work with the other DAs to figure out the ‘route’ to the DataDive weekend.

What did you enjoy the most about being a Data Ambassador, and what was challenging?

Working with people is one of the most enjoyable parts: the charity representatives, other DAs, and the weekend volunteers. It’s amazing seeing them volunteer their time and effort to support the charities.

There were challenging moments over the weekend when we noticed an error in the data, or volunteers asked for something that we hadn’t included. But even if things don’t go exactly to plan, you have people to bounce ideas off and help to resolve problems.

The weekend and everything leading up to it happens in a fairly short period of time. It’s a huge change of pace compared to my job, but very rewarding, because you see the results and impact sooner as well. Sometimes you wonder if you will get enough done over the weekend, or if your skills are ‘good enough’. Volunteers are often surprised at how much they achieve — they are always good enough! Even early on in the DataDive weekend, the charity representatives are saying ‘Wow, this is so useful! We’ve never thought about this; we’ve never been able to do this.’

What advice would you offer to future Data Ambassadors?

You don’t have to be an expert in everything — you work with each others’ strengths. And although it’s a quick turnaround, you have time and support to figure things out. The DataKind staff are supportive, and there are lots of checkpoints, so you aren’t alone.

My advice is to keep an open mind, and don’t be afraid to try new things or learn new things. You have limited time and resources so it’s all about teamwork!

At the moment I’m working with Material Focus (an electronics recycling social enterprise) on a lot of maps-related data, which I haven’t dealt with before. I’ve learnt a lot working with the other Ambassadors — it’s a collaborative environment. By being around them and having conversations, I’m learning so much about how to process geo-related data. I’ve never had a reason to before, so I wouldn’t have known where to start, but now I can see how to approach it.

What is your data background?

At university I studied electronic engineering, and took a module called the Foundations of Machine Learning (ML), so I was introduced to processing data using these techniques. I also studied image processing, which led to my third year project in computer vision. It was part of an agriculture project to build robots that could use visioning to ‘see’ apples, and so ‘pick’ them without damage. My project was focused on identifying apples in images — that’s how I got into it!

I went through a machine learning route, not a statistics route, but I realised that there are data-related aspects to most areas and there’s so much more out there. In electronics we’d build a motion sensor, so the outcome is that we’re processing the sensor data. In a lot of different industries, we’re all using and processing data in a similar way to a data scientist, but for different applications, so we don’t think of it that way. I didn’t really know what data science was until I looked into it!

What kinds of data projects do you currently do, and what tools do you use?

Currently, I’m part of the data science team for SAS, which builds software for data analysis and processing. I do quite a lot of natural language processing (NLP), as well as data analytics and ML. The interesting part about this is that my company has its own language, and drag and drop dashboard tools. I used MATLab in my undergraduate degree, because I had six months to get my head around computer vision and figure out how to implement it. For open source I tend to use Python — I’ve been using Python since secondary school.

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

What I didn’t realise going into data is that being proficient in Python for coding is different to using Python for data science. Nobody told me at the time that you need data-related packages and libraries — that would have made it easier to use Python for data science.

For those looking into data science, there are a lot of different things to develop expertise in, and problems that you could be solving, and it can be overwhelming. For me, it’s better to do something rather than trying to figure out the ‘correct’ thing to do. No matter what you choose to do first, you’ll learn something from it. Then you can use that to decide what your next step is rather than having it all laid out beforehand. Take one step before you try and run a marathon!

Did you always think you were going to go into data?

No! I decided I wanted to do engineering quite early on — I wanted to get into technology in general. When I was younger I wanted to build robots, robots seemed pretty cool. That drew me to engineering, but I also like problem solving, so I saw this as a way to solve problems using technology. When I transitioned to data science, I still viewed it as problem solving, but instead the tools are data to make smarter decisions. I like using technology and innovation, and that what’s data science gave me.

What surprised you most about your volunteering experience?

It surprised me how early on in a DataDive project the charity partners get value from the process. From the outside you see the end result. But during our meetings before the DataDive weekend, preparing the data helped them understand more about the data they do have. Our questions for them about how they were framing something made them think about what to ask of their own data. The process in itself was useful for them.

Is there a resource you’d recommend to the community?

Stanford’s free lectures helped me learn computer vision from scratch.

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

I’m learning to swim — I learnt at school but haven’t done it for a long time, and I’ve been wanting to get better. When I last went on holiday I started doing some swimming again — so I decided I was going to have lessons so that next time I’m on holiday, I’m a pro! It’s fun, and it’s a good form of exercise.