Transforming Data Analysis for a Swiss Commodity Trading House with a Unified Data Stack
article30 Sep 2025
Challenge
The data analysis team at a Swiss commodity trading house faced significant challenges with their existing data infrastructure. Data was scattered across multiple databases in different formats, with no unified schema. The end-of-day data ingestion process took nearly a full day, delaying decision-making. The system also lacked scalability for new data sources and had slower development cycles, limiting adaptability and innovation.
Solution
To address the client's challenges:
- TimeBase unified the trade desk's internal data.
- Custom ETL data loaders processed L1, L2, and L3 data.
- Data preparation was enabled directly from raw files.
- QQL and a Python API provided rich query support for data scientists, traders, and portfolio managers.
- An event-based architecture supported native model streams for events, trades, and orders.
- API and GUI tools facilitated work with high-frequency or irregular interval data.
Results
- The solution enabled complex queries to be prepared on the fly, reducing time to analysis.
- Infrastructure costs and complexity were lowered, while data integrity improved through schema evolution.
- ETL latency was drastically reduced, and scalability allowed seamless addition of new data sources, enhancing user satisfaction.