Improving financial data-driven decision making
Challenge
improve financial decision- making practices.
KLM Airlines wanted to improve its financial decision- making practices. Under EU law, airline financial controllers need to estimate the cost of claims arising from disruptions in advance, so they can make a financial provision: a reserve of money to be paid in the future.
Solution
Value-adding projects.
After completing our month-long Data Science Bootcamp, several Xccelerators joined KLM in its quest to become a more data-driven organization. Working within an existing financial data analytics team with other highly-skilled professionals, our data scientists identified several value-adding projects they could deliver. For example, by building machine learning models with financial data, they could pinpoint fraudulent payments. They could also predict refund amounts and dates for passengers’ claims arising from flight disruptions. An important part of the solution was adding automated improvement functionality through back-testing. A user interface and APIs have been created to make sure that the results can be used by other teams within KLM.
Result
Preprocessing pipelines that are more efficient, complete and unit-tested.
Working with an engineer, a proof of concept (POC) was quickly industrialized and put into production, resulting in preprocessing pipelines that are more efficient, complete and unit-tested. In addition, the entire process is now monitored using predefined KPIs. Plus, other teams within KLM are able to collaborate by making use of the solution that we created. As a result of this initial success, Xccelerated is now collaborating with other Air France-KLM product owners to identify additional value-adding data science projects to develop and implement in the near future.
Other references
