Project

Vehicle Make/Model Standardization

Data cleaning / NLP • Completed

Summary

A standardization pipeline to normalize messy vehicle make and model text for analytics and matching tasks.

Problem

Vehicle datasets contain inconsistent manufacturer and model names that break downstream reporting and matching.

Approach

I built a combination of rule-based normalization and token matching, then validated results against reference data for high coverage.

Results

The solution improved make/model consistency from 60% to 94%, enabling more accurate fleet and insurance reporting.