Goodlookup is an intelligent function designed specifically for spreadsheet users, aimed at solving fuzzy matching and topic clustering problems in data joins. It is based on a pretrained model and combines GPT-3-like intuition with fuzzy-matching join capabilities, and can be used in Google Sheets like vlookup or index match, thereby accelerating topic clustering work.
Main capabilities:
- Semantic matching: Goodlookup can match similar text like a human, recognizing semantic relationships, synonyms, and even cultural similarities. For example, it can match "Ronaldo" to "Ronaldo Luís Nazário de Lima," and it can also match "Ronaldo" to "Soccer" or "Football."
- Fuzzy matching enhancement: The main limitation of traditional fuzzy matching is that it does not consider similarities beyond strings, whereas topic clustering requires semantic understanding. Goodlookup leverages advances in natural language processing to approach semantic understanding and help solve text-to-text record linkage problems.
- Match scoring: The score next to a match result indicates the strength of adjacent words in vector space. Longer strings that exactly match longer strings receive the highest scores.
How to use:
- Subscription: A subscription is required before use, with an annual subscription fee of $15.
- Installation: Install Goodlookup from the Google Workspace Marketplace.
- Activation: In the spreadsheet menu, activate the function via "Extensions" > "Add-ons" > "Goodlookup."
Use cases: Modern spreadsheet users face data join problems, with data scattered across multiple places and naming conventions not unified, making it difficult to obtain a clear and unified view of the data. Goodlookup helps solve this type of text-to-text record linkage problem and is suitable for spreadsheet work that requires topic clustering, data merging, or fuzzy matching.
Limitations: Goodlookup is not intended to replace fuzzy matching, but rather to serve as a new tool in the data manipulation toolbox. Its specific performance limitations, compatibility details, and so on are not yet provided.
