Inspiration

Real estate investments involve much more than just finding a property at the right price. Before purchasing or developing land, investors need to consider environmental hazards, zoning restrictions, ownership history, legal complications, and potential financial risks. Unfortunately, much of this information is scattered across county records, government databases, and other public sources, making the due diligence process complicated and time-consuming.

For the CBRE Challenge at HackUTD 2026, our team wanted to explore how technology could make this process more accessible, efficient, and informative. We were inspired by the idea of connecting fragmented property information into one platform that could help investors and developers identify potential risks before making major financial decisions.

What It Does

Our project is a geospatial real estate intelligence platform designed to bring property information and risk analysis together in one place.

The application is built around an interactive property mapping system that connects geographic parcel boundaries with information about environmental hazards, ownership relationships, legal encumbrances, and development considerations.

Our goal is to help users understand not only what risks may exist for a property, but also how those risks could affect its development potential and financial value.

We focused our initial implementation on Bexar County, Texas, while designing the underlying architecture with future expansion to other regions in mind.

How We Built It

We divided the project into four major components: data ingestion, geospatial infrastructure, risk analysis, and frontend development.

For data ingestion, we used Python to retrieve property records from public geographic information services. We developed an automated pipeline connecting to the Bexar County Appraisal District's ArcGIS API, allowing us to retrieve property identifiers, addresses, and geographic boundaries.

We used Supabase, PostgreSQL, and PostGIS to build our shared geospatial database. PostGIS allowed us to store property boundaries as geographic polygons, validate their geometry, and prepare the data for spatial analysis against environmental hazard datasets.

Our importer supports batch processing, pagination, duplicate prevention, checkpoints, and error handling. During development, we successfully processed 2,000 property records and stored 1,996 valid Bexar County parcels in our database.

We also designed database structures to represent ownership entities, relationships between organizations, and title encumbrances, supporting future legal and ownership risk analysis.

Our broader application architecture incorporates Python-based analytics and a Next.js frontend for presenting property information, geographic data, and risk assessments through an interactive interface.

Challenges We Faced

One of our biggest challenges was working with real-world government data. Although public property information is available, the records are not always consistent or immediately usable.

We encountered properties with missing addresses, invalid identifiers, and geographic boundaries that failed geometry validation. We had to build validation rules that could detect problematic records without interrupting the entire ingestion process.

Another challenge was understanding and implementing geospatial data. We learned how to work with Polygon and MultiPolygon geometries, coordinate reference systems, and PostGIS spatial functionality. Ensuring our geographic data could be stored consistently in PostgreSQL required additional testing and database adjustments.

Integrating multiple components developed by different team members was also challenging. We had to coordinate our database schemas, agree on shared property identifiers, and ensure that our code could work together without interfering with other parts of the application.

With limited time during the hackathon, we also had to balance our original vision of a nationwide platform against the importance of building and validating a functional prototype.

What We Learned

Throughout this project, we gained experience working with technologies that are used in real-world data engineering and geospatial applications.

We learned how to interact with public APIs, build automated data pipelines, manage spatial data in PostgreSQL, and validate large collections of geographic information.

We also developed a better understanding of how complicated real estate due diligence can be and how difficult it is to combine information maintained by different organizations.

Perhaps most importantly, we learned the value of building a reliable foundation before adding more advanced features. Having accurate property data and a consistent database structure is essential before attempting predictive analytics or AI-generated risk assessments.

What's Next

Our long-term goal is to expand the platform beyond Bexar County and support property analysis across the United States.

Future improvements include deeper integration with FEMA flood hazard data, more comprehensive zoning and permit analysis, automated identification of ownership and title complications, and financial models that estimate the potential costs associated with identified risks.

We also hope to expand the platform's AI capabilities to explain how different risk factors may interact, helping users make more informed real estate decisions.

Ultimately, we want to make property due diligence more transparent, connected, and accessible by bringing together geographic data, public records, and intelligent analysis in one platform.

Built With

Share this project:

Updates

Submission history