Xueqiu Influencer Post Tracking

Problem It Solves

Xueqiu user posts are scattered across personal timelines, so manual browsing misses important points and makes it hard to trace or compare views over time. This skill turns “tracking a user” into an engineering workflow: resolve the user_id, fetch statuses/user_timeline.json by date window, then deduplicate, filter, score, and archive posts into auditable files.

How It Works and Boundaries

The main flow is:
- Resolve user ID: parse it from xueqiu.com/u/..., or locate the display name via search when needed.
- Lightweight fetch: prefer direct JSON access with requests or WebFetch, paging through results and applying a date window.
- Pipeline processing: use standard-library scripts for deduplication, low-value filtering, near-duplicate detection, and quality scoring.
- Outputs: generate files such as 总表.csv, 候选全集.csv, 剔除明细.csv, and 质量检查.txt, plus generated summaries of viewpoints and disputed points.

The default is full-timeline mode, ranking by quality score; --company enables company-filter mode, keeping only posts that mention the target company. It fits weekly investor tracking, opinion-drift detection, and methodology mining. Caveats: WAF blocks may require a login cookie or a heavier CDP fallback; it is for timeline recall, not real-time trading signals, and it does not replace source verification.