The second day of the Data Engineering Summit 2024 at Bangalore featured insightful presentations by two esteemed professionals from Booking Holdings: Sucharita Chatterjee, Senior Engineering Leader of Data Engineering, and Ankur Garg, Manager of Data Science Insights. Both speakers provided an in-depth look into the intricate data strategies and engineering marvels at Booking Holdings that facilitate its vision of creating a connected trip experience.
The Elephant Analogy: Understanding Booking’s Scale
Sucharita commenced the presentation with an engaging analogy, comparing Booking’s vast and varied operations to the story of blind men describing an elephant by touching different parts. This analogy aptly illustrates how different departments within Booking perceive and interact with its services.
She explained that while most people associate Booking.com primarily with accommodation bookings, her team’s focus is on other travel services such as flights, car rentals, and airport transfers. Highlighting the massive scale of Booking, Sucharita shared impressive statistics from 2023: 1 billion room nights booked, 28 million property listings, and operations in 220 countries.
Connected Trip Vision and Data Strategy
The core of Booking’s strategy is to make every trip a connected trip, offering a seamless, personalized travel experience. This ambitious vision relies heavily on a robust data strategy comprising:
- Cloud-Based Data Platform: To consolidate legacy data from various acquisitions, ensuring all data is available on a single platform.
- Data and ML Platform: A modern, compliant, and self-service platform for data analytics and machine learning.
- Consumption Patterns: Enabling various data consumption patterns for analytics, machine learning, and experimentation.
Sucharita detailed the architecture of their data platform, emphasizing the importance of data stewardship, governance, and privacy. She highlighted how data is ingested, processed, and consumed across different layers, using technologies like Amazon S3, Snowflake, Apache Spark, and DBT.
Machine Learning and Data Consumption
Ankur took over to delve into the role of data consumers in a modern setup, particularly focusing on machine learning (ML) and analytics. He described how Booking’s ML infrastructure involves creating and storing features, serving model scoring requests, and observing model performance.
He emphasized that data consumers also become data producers, generating new data through their models. This necessitates strict governance, scalability, and the ability to look both backward and forward in data lineage.
Real-Time Analytics and Experimentation
In today’s fast-paced business environment, Ankur pointed out that stakeholders demand near real-time analytics. Booking’s infrastructure supports rapid data processing and metric reporting, ensuring timely insights. He also touched on the importance of experimentation, where hypotheses are tested using historical data, market data, and consistent metric definitions to drive business decisions.
AI Trip Planner: A Glimpse into the Future
Ankur concluded by showcasing Booking’s AI Trip Planner, a feature aimed at enhancing the customer experience by using AI to plan trips seamlessly. This tool combines user input with data from similar customer behaviors to provide optimized travel options. The AI Trip Planner leverages various AI models, including deep learning and generative AI, all supported by robust data infrastructure.
Conclusion
The presentations by Sucharita Chatterjee and Ankur Garg highlighted the sophisticated data strategies at Booking Holdings that underpin its vision of a connected trip. Through meticulous data management, advanced machine learning, and real-time analytics, Booking is poised to deliver unparalleled travel experiences. Their insights at the Data Engineering Summit provided a comprehensive look at how data engineering and analytics drive innovation and efficiency in a global travel giant.
