@ AArete 2025

Context

• Previously, the task requires manually reviewing each incoming data field to determine its equivalent in internal schema, which is time-consuming, error-prone, and limits scalability
• Goal: Create a program to intelligently analyze and match client data to company standard data architecture fields to reduce onboarding time, improve consistency, and enable faster client integration

Final Code Strategies

• Priority First Approach
  • Makes sure fields that are essential to analysis are mapped intelligently
  • Logic is implemented to prioritize these fields throughout the whole process
• Hierarchical Decision Making
  • Exact historical matches
  • Pattern-based rules
  • Fuzzy historical matching
  • General similarity
  • OTHER
• Domain Knowledge Integration
  • Medical abbreviation dictionary
  • Specific mapping rules
• Balancing Competing Goals
  • Uniqueness
  • Exceptions
  • Quality
  • Priority

Process

Final Product

• A Streamlit app where the user inputs test data and priority file, and gets a CSV or XLSX result file