Breaking into Data Science
Top Tips from My First Two Years at the Financial Times
Two years may seem like a long time but in the data science and wider technology industry time flies! For context, I began my career as a data scientist within months of completing university — it was a super fast transition but an extremely exciting one!
In this article, I outline the top tips I would give myself 2 years ago before commencing my role at the Financial Times (FT).
🗣️ Ask Questions
Starting a new role in tech can feel overwhelming and it is impossible to grasp all elements of projects. Whether you are learning a new domain or deepening your knowledge in a familiar one, asking questions is crucial for continuous learning. It shows your curiosity, enthusiasm, and eagerness to be informed. Situations that might prompt questions include navigating a new software platform, mastering the company’s coding best practices, exploring a new machine learning method, or simply identifying the right person to contact within the organisation. These are common scenarios for anyone new to a company or advancing in their career. I have found that asking questions not only helps in understanding the broader organisational functions but also facilitates meeting new people and enhancing technical and soft skills required in my role.
🏔️ Be Open to New Challenges
When presented with challenges it can be nerve-wracking and exciting all at once. However as long as one remains determined and resilient to keep going, the new ventures you encounter do really shape your career.
During my first year at the FT, I was invited to present a data science model to the wider FT organisation with over 100 members in a global audience. It was incredibly daunting and exciting at the same time! Delivering presentations in the workplace has certainly evolved since Covid and the increase of hybrid working. The task provided a chance to develop my presentation skills to ensure the talk was engaging whether the audience was in-person or watching online or watching in retrospect. A key aspect of presenting data science models to ensure a solid delivery is to tailor the content to meet the audience. Following in-depth preparation, the talk was incredibly successful placing the spotlight on the amazing work that data science can contribute for the organisation and towards the achievement of organisational Objective Key Results (OKRs); furthermore this task highlighted to me how much I enjoy contributing to knowledge sharing in the organisation and beyond via giving presentations and answering questions from enthusiastic, passionate and curious audience members.
While the above example shows that challenges can be a solo expedition, challenges may also require external support. Therefore if you find yourself needing additional information or assistance, don’t hesitate to reach out to colleagues and explore documentation from various sources, such as previous projects and established best practices. I have been very lucky to be part of a team which has been incredibly supportive and helped me to become the data scientist I am today!
📝 Clear Documentation & Knowledge Sharing
In the tech world, your path is always evolving: different projects, different teams, different stakeholders. This constantly changing environment necessitates a dedicated effort to consistently share knowledge openly among team members and to create documentation that is clear, concise and detailed. It was quite an eye-opener for me to see the volume of documentation a single project can produce, spanning from the initial scoping phase to the delivery of results.
The list below details the most common forms of documentation you may write or come across in the role and is by no means exhaustive.
- Project Problem Statement
- Technical Scoping Documents
- Sprint Plans & JIRA Tickets
- Project Exploratory Data Analysis
- Code Annotation & Code Comments
- GitHub READMEs
- GitHub Pull Requests / Reviews / Issues
- Ethics Checklists
- Technical Checklists
- Presentations of Results
- Stakeholder Documentation
In all of these cases, the level of technical detail will need to be refined to suit the audience it is intended for. Clear documentation that explains why you are writing it is the most effective type! For instance, why was 0.7 selected as the upper threshold? Why was the date range for the training data limited to the previous 90 days? How were outliers treated — were they removed or encoded? Which customers were included in the model and for what reasons? All these details, while seemingly lengthy or excessive, can be an absolute ‘lifesaver’ to future you or folks who work on the code in the future.
👥 Networking & Collaboration
Jumping into data science is an exciting journey, and it’s important to understand that every project, regardless of its size, is a team effort! In my role as a data scientist, I am frequently working together and in communication with a diverse group of highly experienced and talented individuals, including the wider data science team, stakeholders, software and data engineers, MLOps engineers, product managers and analysts. Given this working environment, it is key to continuously network and collaborate within your data science team but also more widely with other colleagues, who may be involved or benefit from the project you are working on across the organisation. The ability to work with many professionals facilitates an understanding on how various roles perceive data science and also presents a wider opportunity to illustrate the value of data science for the broader organisation.
In my own experience with internal FT collaboration and networking, one significant example was embarking on a secondment following my 2-year milestone. This opportunity involved exploring the FT’s Knowledge Graph and collaborating with the Content & Metadata team, with the outcome of the secondment leading towards stronger connections between the Data Science team and the Content & Metadata team. Additionally, it facilitated the growth and sharing of niche graph data science internal expertise across teams through cross-team presentations and comprehensive documentation, including Cypher Query Language guides. Stay tuned for my upcoming article, where I will delve deeper into Graph Data Science and the FT Knowledge Graph!
Thank you for reading! I hope this article will help in your data science journey!
