At Imperial College Business School, students across our MSc and MBA programmes have the opportunity to get real-world business experience and develop the skills to fuse business technology with data science.
One such opportunity is Data Spark, a scheme jointly developed by Imperial Business Analytics and KPMG, where students get consulting experience with business leaders offering data analytic services. This scheme is now being offered to KPMG clients and other businesses, such as Thomson Reuters and BT Research.
Director of the Centre for Business Analytics, Dr Mark Kennedy, describes Data Sparks as a project that combines data science and consulting skills. He says, “As students demonstrate and extend their skills in projects of significant strategic importance to companies who sponsor the projects, students also learn consulting skills. Our students have been thriving in these challenging assignments!”
Working alongside academic and business mentors, students on the six-week consulting project gain practical business analytics experience and business leaders get access to sought after data analytics expertise. Student participant, and MSc Business Analytics student, Christina Tatli describes it to the Financial Times as a service where businesses can “check-in” data for analysis.
One of the first Data Sparks projects, run in collaboration with KPMG, analysed the commercial applications for a telematics driver app. Consulting a small start-up company with a large telematics data set derived from their mobile phone app, the student team of four students extensively analysed data consisting of 1.4 million driver miles. Analysing driver behaviour before reported crashes, fraud detection and calculating driver safety scores, the team were able to show that a simple mobile phone app could be just as powerful as a black box recorder.
Mazhar Hussain, Director KPMG Digital and Analytics says, “The ability to solve complex business problems using analytics has become a competitive differentiator between average and high-performing organisations. This has made it critical for businesses to understand their data and have the ability interrogate it to make significant business decisions.
“This is forcing organisations to become more data driven in their approach,” Mazhar continues, “There is growing demand for organisations to be able to take a hypothesis about their business, whether it is an opportunity or challenge and test it with agility in a safe environment.”
Since then, Business School students (and some students from across Imperial College) have been providing top companies with data insights and building their skillset in data analytics – a sought after skill across sectors.
Full-Time MBA student George Chatzaras was part of a team of MBA, MSc and MSc Computing Science students who partnered with KPMG and BT Research to identify and analyse cyber-attacks. He says, “It is really a great opportunity to be able work on real life projects with a leading company such as KPMG.”
“It was also very interesting to see how the dynamics of a very diverse team change. A Data Sparks team is usually comprised of computer/software engineers, business analytics students and purely business people,” he continues, “It’s interesting to see how these different team members interact and fill the different gaps according to their talents.”
Students across MBA and MSc programmes at Imperial College Business School can apply for the scheme. In 2017, over 90 students applied and 40 students were awarded places. Some of the Data Spark 2017 projects were:
- Minimising journey time through optimal incident management.
- BT Research and KPMG: understanding Cyber Attacks.
- Royal Bank of Canada: gendered paths to promotion.
- Professional Services of the Future.
- Thomson Reuters: baseline prediction and visualisations of the passage of a bill in parliament.
Drift to Management
A group of students provided key people management insight to the Royal Bank of Canada (RBC) in collaboration with KPMG. The bank had noticed an increase in the number of managers across the organisation, known as “a managerial drift”, and wanted to know the drivers for their promotion and retention. “They knew this was happening,” Christina Tatli told the Financial Times, “but now they know exactly where in the organisation it is taking place, which makes the issue possible to target.”
Robert Carlyle, Senior Director of Strategic Workforce Management at RBC told the Financial Times, “People analytics is an excellent way to identify and drive good performance.”
Team member and Full-Time MBA student Suzy McClintock described it as “no small undertaking in just six weeks… The icing on the cake was of course that both KPMG and the client were thrilled with our work, both the presentation and the final report, with the client very kindly offering to hire us all on the spot!”
Visualising future legislation
Thomson Reuters wanted to understand the data science behind legal data passing through the Government and policy making bodies. The ultimate aim of this team project is to identify and model features involved in the creation of new laws. Using baseline prediction and visualisations, a group of students from MSc Finance, MSc Business Analytics, MSc Management and the Full-Time MBA, successfully identified the likelihood of future laws from 2015-2016 parliamentary sessions.
Using data science techniques, the team provided commercially insightful information to Thomson Reuters on the key factors involved in predicting success outcomes and conducted a sentiment analysis of public opinion and interest to report on the link between news coverage or social media activity and new regulations.
Dr Tharindi Hapuarachchi, Technical Partnerships Manager at Thomson Reuters Labs remarked that they were “impressed with how quickly the students were able to understand and tackle a problem in a new domain, and deliver insights within a few weeks”.
Working with BT Research and KPMG, a team of students from MSc Business Analytics, Full-Time MBA and MSc Computing Science have developed a new way of combatting cyber-attacks. The recent WannaCry ransomware cryptoworm that infected 200,000 computers in more than 150 countries was covered extensively in the media.
Dr David Birch, academic mentor of the team, described the main challenge facing those attempting to identify and put a stop to attacks is the volume of information they have to deal with. “Identifying and analysing cyber-attacks requires sifting through vast amounts of data to find anomalies,” he says, “The team’s work shows the potential of visualisation and user classification to understand and identify anomalies.”
Working with a data set of 18 billion daily events (of which about 0.000075% is cyber-attacks), the team’s breakthrough was to identify standard behaviour between groups of users or “families” and filter it out, leaving a much smaller data set of unexpected behaviour. By visually representing the data, the team was able to map out changes in network connectivity and identify compromised users.
For George the project was “challenging” but, he says, “We are very happy to have contributed to the research and thought-processes behind facing these threats.”