The Ultimate Fuzzy Matching Engine for Data Cleaning & Deduplication ✨
MakFuzz is a powerful, desktop-based fuzzy search utility designed to find needles in haystacks. Whether you are deduplicating customer records, matching messy implementation lists, or just trying to find "John Smith" in a list of "Jon Smythes", MakFuzz has your back!
It goes beyond simple string matching by combining Spelling Similarity (how it looks) with Phonetic Similarity (how it sounds) to give you the most accurate matches possible. 🎧🔡
- Spelling Strategy: Utilizing
Jaro-Winklerdistance to catch typos and visual similarities. - Phonetic Strategy: Powered by
Beider-MorseandDouble Metaphone, it understands how names sound across different languages (French, English, Generic).
- Dynamic Column Selection: Choose any columns from your CSV to use as search criteria—not limited to first/last names!
- Weighted Criteria: Assign different importance to each criterion (e.g., Weight 1 vs Weight 3) to drive your scoring logic.
- Multiple Criteria: Match on as many columns as you need—mix and match names, IDs, addresses, or any data fields.
- Strict Thresholds: Set minimum quality bars per criterion. E.g., "Must be at least 80% phonetically similar to even be considered."
- Live Dashboard: See real-time breakdown of scores for each criterion:
S%(Spelling Score)P%(Phonetic Score)Total Score(Weighted Average)
- Deep Dive Stats: The footer isn't just a footer—it's a data goldmine showing:
- Max Under Threshold: The best match that didn't make the cut (your "near misses").
- Min Above Threshold: The worst match that did make the cut.
- Total: Instant count of all findings.
- Auto-Save: Closing the app? No stress! MakFuzz saves your entire workspace (paths, weights, thresholds, selected columns) to
~/.makfuzz_config.xmlautomatically. It's ready exactly as you left it next time. 🔄
- CSV & Excel: One-click export of your filtered results to standard formats.
- Enhanced Excel Export:
- Color-coded headers (blue for synthesis columns, green for original data)
- Auto-filter enabled on all columns
- Frozen panes to keep search results visible while scrolling
- Auto-sized columns for perfect readability
- Full original row data appended as individual columns
Built with robust, industry-standard libraries:
- Java 17+: The solid LTS foundation.
- Swing + FlatLaf: For a crisp, modern, high-DPI aware user interface.
- Apache Commons Text/Codec: The heavy lifters for string algorithms.
- Jakarta XML Binding (JAXB): For clean, standard configuration management.
- Apache POI: For native Excel (.xlsx) generation.
- Lombok: For cleaner, boilerplate-free code.
-
Build the App:
mvn clean package
(This creates a shaded uber-jar with all dependencies included)
-
Run It:
java -jar makfuzz-1.0.jar
-
Use It:
-
Step 1: Point it to a CSV file (using
,or;separators).[!IMPORTANT] The file must include a header line with column names.
-
Step 2: Select which columns you want to use as search criteria from a dialog.
-
Step 3: Enter search terms for each criterion.
-
Step 4: Adjust weights and thresholds for each criterion to fine-tune your matches.
-
Step 5: Hit Run Search and watch the magic happen! ✨
-
MakFuzz isn't just a tool; it's a productivity booster designed to make data cleansing satisfying. With its clean layout, responsive sorting, and detailed feedback loops, you'll actually enjoy fixing your data.
Happy Fuzzing! 🐰🔍
We're constantly evolving! Here's what we're planning for future releases:
- Auto-Detection: Smart detection of column types and suggested weights.
- Contextual Similarity: Incorporate semantic understanding for even smarter matching.
- Batch Processing: Process multiple files in one go.
Love the tool? Help keep the algorithms fuzzy and the UI crisp! Your support fuels future updates and keeps the caffeine flowing. ☕✨
Every bit helps us build better tools for you. 🚀