
There is still time to visit the Educating for Global Impact (EGI) exhibition on the ground floor of the Marshall Building – until the end of the term. Our MSc Data Science capstone projects have been showcased there! Dr Marcos Barreto goes into detail below about the capstone projects and their aims.
The capstone is a collaborative project, providing students with the opportunity to study and research in depth a data science problem proposed by an industry or academic partner. The problem will normally relate to data-based issues faced in practice by private or public sector organisations, requiring the use of specific data sources and data science skills learnt on the programme. The project is jointly supervised by an LSE faculty and company partner collaborators. It requires creative work in formulating research questions and hypotheses, identifying the most suited methodology, referring to research literature, analysing data sources using a variety of data science tools, and providing insights about the observed results and future directions. The students are able to develop and improve the hard and soft skills needed in most data science jobs.
The list of partners so far includes Microsoft, Tesco, Adobe, Facebook/Meta, Samsung, Alpha Telefónica, AstraZeneca, Capgemini, Koa Health, Siemens Advanta Consulting, Deutsche Bank, Agricom/Huggin Munin, Pareto Economics, ADIA, Google Research, Houghton Street Ventures, Plymouth Marine Laboratory, LSE Planning Division, Department of Methodology (LSE), Department of Government (LSE), Thames Valley Violence Reduction Unit, and The Open University.
The topics covered by the projects span different areas, such as healthcare, finance, social sciences, logistics, agriculture, and environmental science. Some recent examples comprised i) forecasting of sea water transparency and monitoring of water pollution, ii) machine learning models applied to mental health interventions and monitoring of depression symptoms, iii) network/sentiment analysis applied to social media and company network data, iv) prediction models applied to supply chain and reorder point optimization, and v) use of alternative data in financial market forecasting.