case studies
London Stock Exchange
The problem: The London Stock Exchange was facing an issue: they were producing tons of data each day and storing it in local drives. This was expensive, slow, and inflexible. They were about to reach a point where the capacity had to be upgraded again.
The solution: Migrating to Microsoft Azure. Using AzCopy, more than 300 million files were copied to a data lake in the cloud. With a set of Azure Function Apps, metadata was collected from the on-premises server, and the combined data was made available via an API. New data was written directly onto the data lake with an adjacent write API.
The effect: An easily scalable, much cheaper solution to support data storage needs in the long run.
Copenhagen Infrastructure Partners
The problem: Copenhagen Infrastructure Partners was using complex Excel-based systems to produce financial and environmental reports, which were intransparent and prone to human error. This was affecting deadlines for report submission and confidence in the values.
The solution: A Python-based data platform built in Microsoft Fabric to centralize and organize the data sources, as well as perform cleaning, transformations, and apply business logic. Thanks to integrated data sources based on Logic Apps, this automated the reporting process and allowed for easy further integration with BI tools. This was coupled with extensive documentation.
The effect: Timely submission of reports and easy access to all environmental and financial data. Automated reporting for long-term gain and simple external audit processes thanks to the documentation.
Booking.com
The problem: After several mergers and acquisitions, the geographical database supporting Booking.com and its services was full of duplicates and otherwise misaligned with the UX vision. Locations were inconsistently defined and available in multiple versions, causing confusion to the users.
The solution: A thorough cleaning of the geographical database assisted by machine learning to identify duplicate values. In addition, a development of a convention for the definition and naming of locations and its subsequent implementation.
The effect: A significant improvement in the usability of the location search in Booking.com and related services, and consistency across the geographical database.
World Bank
The problem: Each year, World Bank would conduct an analysis of real-estate markets in which it was present. This required hundreds of emails with requests for information about individual markets, which was then manually converted into Excel spreadsheets. The collected data was visualized in a PowerPoint presentation. This was slow and inconsistent in format, leading to incorrect data being acted upon or data not being comparable across years.
The solution: The data collection process was moved to an integrated survey system, which contains comprehensive input validation to ensure consistency of format. This was fed into a Power BI dashboard with an intuitive visualization across markets and years, with adjustable granularity.
The effect: Constant access to updated data, easy comparison between markets and across time, and a major improvement the quality of the data - and, what follows, the quality of the business decisions.
JLL
The problem: JLL, a global consultancy in the real-estate management industry, was dealing with a complex web of information. Occupancy data, building plans, and more was unpredictably scattered across the organization, leading to very slow and inconsistent planning.
The solution: Using Azure Databricks, the data was consolidated in a single place, serving as a single source of truth for more than 20 major accounts. A single comprehensive Power BI dashboard was created to visualize all real-estate metrics, designed in close cooperation with the business end of the company.
The effect: Major efficiency gains, assisted by the availability of current data in a predictable format. In addition, the standardization produced by centralization also ensured consistency across the organization.