Overview
Anyone who wants to buy or sell a plot of land in Germany faces the challenge of determining a realistic price. Publicly available standard land values (BORIS data) provide a basis, but they are often outdated, incomplete, or hard to interpret. For a leading real estate portal, I developed a microservice that closes this gap.
The land price calculator combines BORIS data with proprietary data collection into a proprietary calculation model that computes current land prices by city and region in Germany. The service was designed as a standalone Django microservice and made available via a REST API.
Initially, the platform used the API internally to automatically generate data-driven landing pages for plots of land in German cities. Following the success of this approach, the API was gradually opened up to external users as well, evolving into a standalone product.
My Role
As CTO and developer, I was responsible for the project from September 2023 to April 2024, from the concept phase to production operation. This included both the technical architecture and implementation as well as alignment with the product team regarding the data model and API design. I was responsible for the entire development of the microservice — from the data model and calculation logic to the API layer.
Tasks
- Design and development of the proprietary price calculation model — Designing the calculation logic based on public and proprietary data sources, with the goal of the most precise estimates possible at the city level
- Preparation and integration of external and proprietary data sources — Processing BORIS data from different federal states and integrating proprietary data collection into a unified data model
- Development of the Django REST API — Providing structured endpoints for internal systems (landing page generation) and external consumers
- Automated generation of data-driven landing pages — Connecting to the platform’s content system for automatic creation of land pages for German cities
- Gradual opening of the API to external users — Introduction of authentication, rate limiting, and documentation for external access
Technical Highlights
Processing heterogeneous BORIS data: In Germany, standard land value data is provided at the state level — in different formats and with varying data quality. A central challenge was normalizing these heterogeneous sources into a unified, queryable data model in PostgreSQL.
Proprietary price calculation model: The calculation model goes beyond simply reproducing standard land values. By combining public BORIS data with proprietary data collection, a more nuanced price estimate emerges that takes regional characteristics and current market developments into account.
API design for the transition from internal to external: The API was designed from the start so that it could initially be consumed internally without access restrictions, but could also be opened to external users with minimal adjustments — authentication, rate limiting, versioned endpoints.