Exploring Fuzzy Matching with Python

Fuzzy matching is a technique used to identify similar but not identical text entries, particularly helpful when handling misspellings or formatting inconsistencies. In a baseball dataset, fuzzy matching was used to align player names across two data sources where names like “Gregg Zau” and “Gregg Zaun” referred to the same player. A custom function applied the fuzz.ratio method to compute similarity scores and match player names above a set threshold. The same approach was extended to match stock listings between U.S. and U.K. exchanges, where company names were often listed slightly differently. Fuzzy matching successfully paired entries like “APPLE INC.” on NASDAQ with its equivalent on the London Stock Exchange. To improve accuracy, only matches with similarity scores of 90 or higher were retained. Matched records were then merged and compared by stock price using Yahoo Finance data. Several stock price discrepancies were observed, likely due to currency differences, market conditions, or the use of depositary receipts. The project also demonstrated fuzzy comparisons on names and addresses, showing varying degrees of similarity. Finally, common fuzzy matching algorithms include Levenshtein distance, Jaccard similarity, and cosine similarity, all of which support flexible, real-world data cleaning and integration tasks.