What does it take for a data science team to move from statistics and model design to actually delivering value within a corporate environment, where customers have very different backgrounds and requirements?
With 500,000 flights and over 80 million passengers a year, internal teams specialising in operations, maintenance, scheduling, network expansion, pricing, consumer preferences, marketing and inflight retail, and a growing data science team, easyJet is a representative example of a diverse environment with a lots of opportunities and even more potential traps for aspiring disruptors.
Some of these challenges include:
- Business problems require a variety of approaches/techniques to solve, often extending beyond core data science skills (e.g. operations research, web development). Expert support from IT and other functions is often unavailable.
- Data scientist turnover can lead to projects being discontinued.
- Business teams have their own legacy systems that they swear by and scepticism or resistance to change is common.
- Business functions are often not fully aware of what they want to achieve when engaging the data science team.
- Data sources are distributed/isolated or even completely inaccessible.
- Bringing models into production can mean different things and there is a trade-off between what is useable and what is fast to create.
In this session we will go through a number of projects from easyJet’s recent past and give practical advice on how we solved these challenges and how we have adapted our processes, structure, collaboration practices and infrastructure, in order to avoid many of these in the future. The projects we will touch upon are:
- Predictive maintenance: We worked alongside the engineering team to propose optimal replacement times for the thousands of electrical and mechanical components found in an Airbus aircraft.
- Slot matching: We worked with the scheduling team to create an interactive tool that matches our commercial schedule with airport slots, given user constraints and costs.
- Food waste: We collaborated with the inflight team to minimise food waste, based on machine learning techniques.
- OTP forecast: We built a model to predict when flights arrive on time and inform operations to take pre-emptive action, using customised visualisations.
- Pricing: We have developed a number of ML algorithms to optimise pricing decisions and augment the capabilities of human pricing managers.
Alberto Rey is head of data science at easyJet, where he leads easyJet’s efforts to adopt advance analytics within different areas of the business. Alberto’s background is in air transport and economics, and he has more than 15 years’ experience in the air travel industry. Alberto started his career in advanced analytics as a member of the Pricing and Revenue Management team at easyJet, working in the development of one of the most advanced pricing engines within the industry, where his team pioneered the implementation of machine-learning techniques to drive pricing. He holds an MSc in data mining and an MBA from Cranfield University.
Grigorios has more than two years experience as a data scientist in easyJet. He has applied machine learning and statistical techniques to tackle a big variety of business problems and drive profits and savings within the company. He enjoys working with customers and contributing to the evolution and growth of his team. He holds a PhD in Electronic and Electrical Engineering from Imperial College and has an interdisciplinary background in Bayesian modelling and parallel computing.
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
©2018, O’Reilly UK Ltd • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • email@example.com