Identifying Your Ideal Data Science Role

Identifying-your-ideal-data-science-role-625x350

 

It’s fair to say that one of the characteristics of most professional occupations in the 21st century is that there is no such thing as a “typical” day. This is no less true in the world of data science. Nonetheless, different types of organisations can have very different expectations of their data scientists which, to some extent, you can anticipate.

The essence of data science is to analyse data with the goal of producing insight. If you were to look at how data scientists spend their time over an extended period, you’d find that they typically perform (at least!) five different activities to achieve this goal. 

  1. Data preparation. Never underestimate how much of time is spent in data cleansing, integration, labelling, etc. This can take up to two thirds or more of the time spent delivering a data science project.
  2. Model building and testing. Developing, training, testing, evaluating and iterating to deliver a data science model with adequate predictive accuracy.
  3. Team collaboration. Working with your colleagues to prioritise projects, co-working on specific data science problems, lunching and learning from one another.
  4. Interfacing with a range of stakeholders: the business users of your insights to specify the problem and deliverables, with the data management team to understand data availability and access, with IT’s software engineers to facilitate deployment into production and post-deployment monitoring
  5. R&D and learning. Keeping abreast of the latest developments in machine learning through social media updates, research papers, conference attendance, etc.

The relative mix depends on the organisation and its data science setup. As you think about your own skills, its helpful to try to anticipate which organisations are likely to offer a mix of activities that will best match your preferences:

  • Maturity of data environment. Mature data management environments (extent of data capture and integration, flexibility of data interfaces, the quality of the data itself) mean less time spent on data wrangling. Don’t forget that, for all their investment dollars, some of the largest incumbent corporations often have more challenging data environments than smaller start-ups.
  • Size of data science team. The larger the team, the more scope for role specialisation. In a large team, junior data scientists may spend 100% of their time in data preparation. There may even be dedicated roles for internal client management. In a very small team, you will have scope to take on all of the above roles and more.
  • Organisational home. A data science team that sits at the Group level astride multiple business units and corporate functions is going to be tackling a much wider range of problems than one that sits within a specific business unit or function (e.g., HR, Finance, Product Development). 

Some organisations have a greater need for multi-skilled individuals with strong interpersonal skills, who are good at multi-tasking and who enjoy variety. Others will be looking for individuals who want to play more to a single area of expertise. 

Whether you’re starting your career in data science or considering moving from your current organisation, have a think about what an ideal role would look like for you and where you’re most likely to find it.

 

 

Back to article list