By Marisa Reventos, Senior Manager – Experiential Learning
LondonLAB is a 10-week consulting style project designed to help an organisation address a current analytical business challenge. This is a faculty-led, experiential course that allows Masters in Analytics and Management (MAM) students to engage in a live project and tackle a business challenge for leading companies in London. Students apply the learning from their core courses to develop solutions that have real-world impact.
Marco Laube (MAM2021) shares his experience of participating in LondonLAB.
Eleven of my classmates and I had the chance to work with the European Bank for Reconstruction and Development (EBRD). Since its inception in 1991, the EBRD has invested almost €150 billion in a total of more than 6,000 projects. The bank is committed to championing progress towards market-oriented economies and the promotion of private and entrepreneurial initiative.
It was extremely rewarding to work with a client with such a significant impact on the lives of millions of people. I was particularly intrigued by the bank’s commitment to promoting environmentally sound and sustainable development.
The Business Challenge
As a multilateral development bank, the EBRD has over 60 offices in more than 40 countries. This results in a mobile and diverse workforce.
Therefore, the client sought to understand how mobility and geographical location have affected the careers of employees; and if there are characteristics of individual persons or organisational units that could explain different levels of geographical mobility or career progression.
To answer these questions, we were provided with a dataset containing anonymised information on more than 6,500 employees covering a period of 5 years.
From the very beginning, we students were in full charge of the project. This included scheduling the meetings with the client, scoping and structuring the project, and defining roles and responsibilities within the team.
We were organised into two groups of 6 students, with each team working on the same research question from a different perspective.
Due to the ongoing COVID-19 pandemic, team members were spread around the world and across different time zones. Consequently, one of the biggest challenges and a great learning was managing and motivating a global team in a fast-paced, dynamic, and virtual environment.
What adds to this is that group members had diverse backgrounds with unique skillsets. Thanks to our core course on Performing in Organisations, we were able to leverage this fact and maximise the quality of our work and the impact on our client.
The Data Science Life Cycle
By far, the best part of LondonLAB is to tackle a real-world problem and to have the chance to execute the entire “Data Science Life Cycle”. You really get a flavour of how it is to work as a data scientist.
Phase 1: Business Understanding
We started by gaining a deep understanding of the business case and had intensive talks with the client. This was of utmost importance, as our project required a careful definition of the objects of analysis. As an example, we had to agree on how to measure “career progression” and how to aggregate this for multiple employees and years. Only when you truly understand the business challenge and know what to analyse can you proceed to the next phase!
Phase 2: Data Preparation
Working on a real-world-problem means working with real-world data. This in turn means that a significant amount of time is needed to make sense of the data and to clean it. We had to join multiple datasets, impute missing values with custom-made algorithms, normalise attributes, and engineer new features. You know what they say: “Data scientists spend 80% of their time cleaning and manipulating data and only 20% actually analysing it” – well, it’s true!
Phase 3: Exploratory Data Analysis
Intertwined with phase 2 was the exploratory analysis of our data. We created over 200 graphs and tables to get to know the data and facilitate the modelling process. This was a vital step as it enabled us to discover and fix systematic data errors that had occurred during the export of the data from the client’s systems and would have otherwise distorted our results.
Phase 4: Modelling
We built over 70 statistical models and were able to apply the whole range of tools and techniques that we had been taught in core courses such as Applied Statistics, Data Science for Business, and Machine Learning for Big Data.
We were thrilled to be supported by our professor Francesca Franco, who provided invaluable inspiration and helped us shape the path ahead.
There was always a need to adapt to new findings as new additional avenues for further investigation evolved. The project was highly dynamic, and we had to prioritise different actions and analyses. This is also why we were in regular contact with the client and the other team to keep aligned.
Phase 5: Explanation and Visualisation
The ultimate goal was to provide insightful analyses and help the client to gain new perspectives. For this, we had to present our findings in a clear way, taking into account the audience’s background. Therefore, we produced various graphs visualising our results, ranging from world maps and alluvial diagrams to well-known bar charts.
It was amazing to see how much work both teams had put into the project, with the final results exceeding the client’s expectations by far.
While we found that there are stark differences in the levels of employee mobility within the organisation, both teams came to the conclusion that this is not due to structural inequalities, but can be explained by the growth of certain divisions and office locations. That is, the EBRD offers equal opportunities across the organisational structure and no systematic discrimination is happening with regards to career progression.
This is an invaluable insight that will inform future decisions across the whole human resources department. Some of the slides were directly used by the People Analytics Team in a presentation to the board of the bank.
All in all, LondonLAB was a transformative experience, with the opportunity to apply all our knowledge and learned skills in one big project, having a real-world impact!