Leading European Airline Benefits from Machine Learning
Codec worked in partnership with an airline to create and design a customized Azure App to pull data from the airline’s existing on-site database to an Azure SQL database. Various data on stock levels, flight schedule/routes and other statistics were fed into the newly-created cloud-based infrastructure whereupon the magic dust of of Azure Machine Learning predictive models was sprinkled on this existing data.
Company: Leading European Airline
Have you ever been on a flight and licking your lips in anticipation at getting stuck into the all-day breakfast burrito only for the food trolley to come your way and an air steward to politely inform you the menu item you ordered was no longer available? Frustrating, right? Well, a leading European airline share your pain and intend doing something about it. That’s why they contacted Codec looking to avail of the world of possibilities and new opportunities presented by advanced Business Intelligence solutions and machine learning.
The reality of course is that an aircraft can only carry a certain amount of stock, so they’ll often run out of a specific stock item (our burrito!), while running a surplus on other items. Therefore the business leadership there expressed a need to project flight sales for future flights to deliver more accurate supply planning data and that’s where Codec stepped in to help.
The objective of the Inflight Sales project that Codec worked on deploying for this particular airline was to analyse historic sales data, and use machine learning to create predictive models. This would help the airline to understand the optimal stock quantities to have on any aircraft or flight route within a small margin of error (10 – 15%).
One of Europe’s leading airlines, carrying millions of customers per annum on regular daily flights from multiple bases across Europe, connecting passengers to urban destinations on a fleet of hundreds of Boeing aircraft. The airline has a team of thousands of highly skilled aviation professionals and a highly impressive safety record.
Codec worked in partnership with the airline to create and design a customized Azure App to pull data from the airline’s existing on-site database to an Azure SQL database. Various data on stock levels, flight schedule/routes and other statistics were fed into the newly-created cloud-based infrastructure whereupon the magic dust of of Azure Machine Learning predictive models was sprinkled on this existing data. The powerful new data models were made available for reporting in Power BI (enabling visualization of the enhanced data from beautiful reports to deliver radical new and informed business insights).
The airline now has the capacity (via the constantly updated reports in Power BI) to visualize and predict stock demand for inflight sales for their flights. The days of passengers moaning about not getting their burrito may be coming to an end. Furthermore as the airline allows more and more data to be accessed by the Azure AQL database, and as more machine learning algorithms are allowed to do their thing, the potential opportunities provided by adding new prediction models to the mix looks to be a genuinely transformative proposition for the company in question.
The airline could conceivably check on the sustainability of prospective new flight routes or perhaps even predict which flights were more likely to be overbooked (preventing the kind of PR disaster that affected United Airlines as a passenger was forcibly dragged from the plane). If you’re not sensing the massive potential of machine learning by now then you’re probably one of those who thought the internet was a passing fad when it emerged in the early 90’s. The question is not so much if this is a viable concept for your business anymore, but rather if you can afford to ignore it any longer.