What does a data scientist actually do?

By Kimberly Cook |Email | May 23, 2018 | 8397 Views

In this feature, we take a look inside the working lives of people whose job titles often warrant the question: 'but what do you actually do?' This week, we speak to Glenn Bunker, data scientist at realestate.com.au.

What do you actually do?
A data scientist's job description can be very different depending on who you speak to. For me, the fundamental thing that defines a data scientist is the ability to identify and solve opportunities and problems and help make objective decisions based on data.

Day to day, I manage a team of data scientists whose purpose is to understand our customers so that we can provide them with the most relevant, personalised and timely experience possible. For realestate.com.au, this includes things like recommendations for someone's next best property, providing an update on the property market specific to someone's property, or providing tips to first home buyers or renters.

To provide these insights, my team performs a range of different tasks on any given day, including:

  • data wrangling, analysis and visualisation;
  • developing machine learning algorithms and building predictive models;
  • testing, monitoring and optimising the performance of models and algorithms;
  • presenting results and recommendations to a range of stakeholders;
  • engaging with product owners across the business to drive future data science opportunities.

We're able to analyse billions of behavioural data points, generated by the more than a million Australians visiting realestate.com.au every day who are engaging with hundreds of thousands of property listings. For example, we'll look at things like, how people are searching for property, what they're looking at, and the properties they're saving.

What does a good working day look like?
A good working day for a data scientist is one where the focus is on exploring data, rather than gathering it. It typically involves learning something new, and includes team collaboration. It involves finding insights from data that are truly valued, and even more importantly, seeing them used effectively by the business.

Recently we put into market a campaign targeting very recent home buyers, identified by a model that our team built, which resulted in our best performing marketing communications ever. In true realestate.com.au tradition we celebrated with cake and champagne. That was definitely a good day.

What's the most difficult part of your job?
Like any data scientist, the part of the job we like least is battling to get access to all the data and making sure that it's clean and of high quality. This is particularly important because a key part of our role is to understand the points of data that are useful in solving a given problem. In practice this isn't always easy. Luckily, at realestate.com.au we not only already do this well, we're getting even better.

What's your favourite part about your role?
Harvard Business Review named data scientist as the 'sexiest job of the 21st century', and with good reason. The explosion of data through digitalisation combined with the vast improvements in technology to store, process and analyse this data, means there are so many great opportunities to explore in this space.

This is certainly true for realestate.com.au, where data is a core part of our strategy and future. For me, my favourite part of the job is the variety of projects we get to work on; from modelling customer behaviours to enhance their experiences, to understanding image content to better understand properties.

We get to analyse all sorts of different data, applying a range of different data science techniques, to extract insight and ultimately add value to the business.

Our renowned hackathons (REAio) also provide a great opportunity to foster innovation, allowing us to test the waters with cutting edge techniques and technologies.

How does your role keep you on your toes?
Data science is a very dynamic space, technology and analytic techniques are evolving rapidly. The ability to analyse data of various formats such as text, images and video has never been easier. This presents all sorts of opportunities, which is very exciting, and keeps us on our toes.

What makes a good data scientist?
Good data scientists are curious, analytical thinkers who are purpose driven and focus on extracting valuable insight from data. They can identify the appropriate data, tools and techniques for a given problem. They know when 80 percent is enough, and when it's not. They understand the business and can communicate complex things in a simple and meaningful way with a range of people at various levels. Finally, they have a thirst to learn; a must in the dynamic, evolving world of data science.


The article was originally published here

Source: HOB