The original aims of the placement were met and exceeded. Key activities initially were multiple online courses on Google Cloud platform, git, and Google BiqQuery. All skills I’m glad to know more about as they are foundational for many data science roles. Then I was given a period of time to get to grips with the data I would be working with and to have discussions with various team members and stakeholders to understand expectations and build a conceptual understanding of what was possible.

Then I was given great ownership over the proof of concept research and was empowered to supplement the code development with in depth research into the branches of machine learning I felt might supplement the work. I iteratively built an additive machine learning model which performed far better than was initially expected and presented that to the technical team.

Following positive feedback and some advisory sessions with some of their in-house experts, I developed it further to better encode geography and to now-cast. I am very proud of the outcome and I understand BT are now looking to publish a paper and scale the work I carried out. I gave multiple presentations to a variety of audiences and received guidance throughout on expectations so felt I learnt a lot. My supervisor had a lot of time for me and allowed a fairly exploratory process which let me learn a great deal in the construction of my model.

The experience has definitely made me a better data scientist. My only advice to future students is to not be intimidated by an overwhelming project brief because by the end of the 3 months I had exceeded everyone’s expectations, including my own.