Comparatory is a web application that provides novel insights about public companies based on free text data from regulatory filings. It uses deep learning to understand a company's business model, and provides several channels to explore the competitive landscape.
Start using the Comparatory web application for free today, or host this open source Comparatory API following the usage instructions below. API specifications can be found in api.raml (rendered into interactive HTML here).
- Set environment variables:
CONFIG: configuration type (config.ProductionConfig,config.DevelopmentConfig)ES_HOST=: URL of ElasticSearch serverAWS_ACCESS_KEY_ID: AWS ID for downloading pretrained modelAWS_SECRET_ACCESS_KEY: AWS key for downloading pretrained model
- Initialize ElasticSearch (one time) with company descriptions:
python index_descriptions.py - Set up virtualenv:
virtualenv venv && source venv/bin/activate - Install requirements:
pip install -r requirements.txt - Run web application:
python run.pyorgunicorn app:app - Profit:
curl -G -XGET 'localhost:5000/companies/describe' --data-urlencode 'description=commercial insurance brokerage'