Geekbi Temu Data Analysis and Market Research

Problem Addressed

Temu product selection and market research often break down when data definitions are inconsistent: products, stores, categories, keywords, and reviews come from separate search paths, and image-search results may be limited to the first page, making conclusions hard to verify. This skill consolidates those queries into a single business entry point and uses Geekbi's returned data to move from candidate product pools to competitor and review validation.

How It Works

The skill routes user intent to internal capabilities: product search, image-search matches, store search, category search, keyword search, and review search. A typical flow is: confirm site and time scope → build minimal query conditions → run the matching script → check pagination and missing fields → output conclusions. Full selection work can follow category → keyword → product → review → store; image-based research first builds a visual candidate pool, then filters by sales, price, match count, and listing time. Output preserves site, time range, sample size, and update time, separating interface facts, recomputable metrics, and analytical judgment. Non-empty linkUrl values are shown as clickable names.

Boundaries and Notes

This skill is for Temu business data, not developer integration or raw API debugging. If no site is specified, it defaults to the US site; cross-site, cross-currency, or cross-period comparisons stop direct ranking. Values of 0, incomplete pagination, or missing fields are explicitly flagged rather than automatically treated as real zero or out of stock.