Fuzzy Matching for Content Gap Analysis & Competitive Opportunity Identification

Content gap analysis typically requires manual comparison of hundreds of URLs—this intelligent Google Colab notebook automates competitive content analysis using fuzzy string matching with RapidFuzz to identify which competitor content pieces you already cover versus genuine content gaps representing new opportunities. Created by Lazarina Stoy for her Introduction to ML for SEOs course, this workflow goes beyond simple exact matching by using URL slug extraction and semantic similarity scoring to recognize when your article about “complete guide on scikit-learn for data science” matches a competitor’s “data science machine learning guide” even when the URLs and titles differ substantially—something impossible with traditional URL comparison tools.
The notebook implements a sophisticated matching pipeline with smart URL preprocessing. Rather than naively comparing full URLs (which would miss matches due to different domain structures, date folders, or category paths), it extracts only the content slug—the final meaningful segment of the URL path—and combines it with the page title to create a rich matching signature. For each competitor URL, the algorithm finds the best match in your content inventory using token_set_ratio scoring (which handles word order differences and partial matches better than simple string comparison), assigns a similarity score, and categorizes it based on a configurable threshold (default 50/100). URLs scoring above the threshold go into “similar_matches.csv” with your matching content identified, while those below threshold go into “missed_opportunities.csv”—critically, this file includes the top 3 closest matches with scores even for “misses,” helping you decide if they’re true gaps or just low-confidence matches you should manually review. The visualization suite includes boxplots showing similarity score distributions for missed opportunities, a Venn diagram (930 similar matches vs 258 missed opportunities in the example), and three-panel n-gram analysis showing the most common shared bigrams, trigrams, and four-grams between similar content—revealing thematic overlaps like “data science,” “machine learning,” “neural networks.”
Use this for:
‧ Automated content gap analysis at scale by comparing hundreds or thousands of competitor URLs against your content inventory without manual review
‧ Identifying content opportunities by isolating competitor pieces that have no equivalent in your content strategy (the “missed opportunities” file)
‧ Validating content overlap by reviewing the “similar matches” to ensure you’re actually covering the same topics competitors rank for
‧ Prioritizing content creation using the top 3 closest matches with scores to distinguish high-priority gaps (low scores = truly unique competitor content) from edge cases (medium scores = maybe related?)
‧ Thematic analysis of competitive content by examining shared n-grams across similar matches to understand common topic framing and terminology
‧ Quality control for content audits by identifying potential duplicate or cannibalized content in your own inventory through self-comparison
‧ Strategic planning by visualizing the proportion of competitor content you’ve covered (similar matches) versus genuine gaps (missed opportunities)
This is perfect for SEO strategists, content marketers, and competitive intelligence analysts who need to systematically identify content gaps across large competitor content libraries—particularly valuable when manually comparing hundreds of articles is impractical and you need data-driven prioritization of which competitor topics represent genuine opportunities versus content you already effectively cover with different URL structures or phrasing.