From start-up to taming big data.
Dynamic data duo Alex & Stephen went from developers for a tech start-up to Data Scientists at a mass media and telecommunications company in the UK.
After a short spell as Python developers at a tech start-up, Alex & Stephen made the step up to the data science team of the UK's largest mobile and broadband network providers to assist their model building processes.
Reporting directly to a team of senior data scientists, both were tasked with creating two machine learning models for a marketing campaign to target specific customers who are most likely to purchase a service from the client.
With such a large subscriber base, one of the main challenges the client faced was the sheer magnitude of datasets, meaning running scripts would take significant periods of time, leading to efficiency issues.
Alex’s responsibility was to build mathematical models that helped different stakeholders within the business make informed decisions, a process that involved manipulating and building large datasets and evaluating and presenting the results to the wider team.
Stephen was tasked with how the client can better target their campaigns to the appropriate customers and increase engagements using historical data. This was done by creating machine learning models using extremely large volumes of data.
Alex & Stephen's responsibilities included:
Communicating with key stakeholders to clarify the requirements and ensure that all criteria was met for the model.
Testing different optimisation approaches within Python and SQL and gather data using SQL queries to build the training dataset.
Used Hyper Parameter Tuning of the model using a grid search, to account for over/underfitting.
Utilising engineering techniques to identify the most important features for the models to reduce the size datasets.
Clean the data in Jupyter Notebook and explain the distribution of the customer base to stakeholders.
Once the model was created, they had to evaluate the performance of the model by looking at the precision, recall, and F2 score.
The models built by Alex and Stephen helped senior stakeholders to make informed decisions and from this, the marketing department used the model outputs to target and retain customers occupying the subscriber base.
Different model metrics were used to evaluate the performance of the model and compared to previous versions of the model.
Stephen and Alex were delighted to learn that their model performed extremely well in comparison to previous models, leading to the client approving its industrialisation.
With the model now automated, stakeholders can now see the model results on a regular basis.
Thanks to the Xander's training I was confident in tackling any problem given to me as they had given me the tools for success, from stakeholder interaction and presentation skills to technical skills like Python and SQL. If there were any areas that I was unsure of, Xander were more than happy to provide me further training that would help me further."
Stephen, Data Science Associate
Xander supported me on this placement by providing upskilling training prior to my start date. I also had meetings with Associates currently placed with the client to gain a better understanding of the culture within the team. I had regular catch-up meetings with Xander during my placement, which made me feel supported and valued."
Alex, Data Science Associate
Want to know more about how Xander can support your business needs?
You may also be interested in...
Our Associates have provided specialist tech and transformation support for startups, charities and the world's largest organisations.