Selected work / 01 · Data integration
From property reports
to structured data.
Commercial Property Deal Report Extraction: bringing semi-structured HTML documents into a consistent, analytics-ready format.
The context
Useful information.
An unstructured starting point.
Commercial-property deal information was held in semi-structured HTML reports in Azure Blob Storage. Making that information available for downstream analysis meant converting documents into a standardized database structure.
At Nimbus Property Systems, I built an extraction workflow for 5,288 HTML deal reports, mapping their content into 27 standardized commercial-property fields and loading structured outputs into PostgreSQL.
The workflow
From cloud storage to SQL.
Read the source documents
Access the HTML deal-report collection in Azure Blob Storage.
Extract the report content
Parse semi-structured documents with Python as part of the extraction workflow.
Map to a consistent structure
Organize commercial-property information into 27 standardized fields.
Load for downstream use
Store structured outputs in PostgreSQL for further analysis.
My contribution
Connecting documents
with the data model.
- Built the Azure Blob Storage extraction workflow for the report collection.
- Parsed semi-structured HTML into standardized commercial-property attributes.
- Loaded the structured outputs into PostgreSQL to support downstream analysis.
The delivery
A consistent format for analysis.
The workflow connected document storage with structured database output. Its scope was a 5,288-report collection and a 27-field commercial-property structure.
The engineering goal: make document-held information usable as structured data.