This project processes public records from Dallas to extract mortgage principal amounts using OCR (Optical Character Recognition). It includes scripts for data scraping, image processing, and visualization.
- data/: Contains parcel shapefiles and the main database.
- images/: Downloaded images used for OCR.
- scripts/: Python scripts for data processing.
- results/: Output files including CSVs and videos.
- tests/: Test scripts to verify functionality.
- audio/: Audio files included in the repository.
To set up the environment, follow these steps:
-
Create the conda environment:
conda env create -f environment.yml
-
Activate the environment:
conda activate deed_ocr_dallas
- Initialize the database:
python scripts/check_db.py- Run the main script:
python scripts/find_principal_dallas.py- Visualize the results:
python scripts/visualize_principal.pyContributions are welcome. Please fork the repository and submit a pull request.
##About This script collects Dallas' parcel-level mortgage data in reverse chronological order. To collect all data since 2020 would require ~$3000 in cloud compute costs or 1 month of compute time on a single computer of the specs below.
This project is licensed under the GPLv3+ License - see the LICENSE file for details.
- CPU: x86_64
- RAM: 15.46 GB
- Storage: 1006.85 GB
- OS: Linux 5.15.146.1-microsoft-standard-WSL2