| name | cn-institutional-survey-heat |
| description | Use when the user asks which A-share stocks, sectors, industries, or themes have the most institutional research visits / 机构调研 heat over the last 7 days, recent weeks, May-June, two months, or a custom date range. Produces stock Top 10, sector/industry Top 10, and weekly sector heat trends from Eastmoney institutional survey detail records. |
| metadata | {"short-description":"A-share institutional survey heat and weekly sector trends"} |
CN Institutional Survey Heat
Use this skill for A-share institutional research / 机构调研 heat questions such as:
最近7天调研最多的股票
最近7天机构调研最多的板块
5-6月每周板块调研热度
哪些行业机构调研在升温
机构调研热度 Top 10
This skill is a data-collection and aggregation helper. It does not give buy/sell instructions. Treat high institutional survey heat as attention and research intensity, not as a trading signal by itself.
Default Workflow
Run the bundled script from the repository root:
python3 skills/cn-institutional-survey-heat/scripts/institutional_survey_heat.py
Default output:
调研最多股票 Top 10: last 7 natural days.
板块/行业调研热度 Top 10: last 7 natural days.
Top 板块周度调研热度: last 2 calendar months including the end date's month.
Useful options:
python3 skills/cn-institutional-survey-heat/scripts/institutional_survey_heat.py --end-date 2026-06-30
python3 skills/cn-institutional-survey-heat/scripts/institutional_survey_heat.py --days 7 --months 2 --top 10
python3 skills/cn-institutional-survey-heat/scripts/institutional_survey_heat.py --format json
python3 skills/cn-institutional-survey-heat/scripts/institutional_survey_heat.py --save-csv --output-dir /private/tmp/survey_heat
Data And Rate Discipline
- Source table: Eastmoney institutional survey detail records,
RPT_ORG_SURVEYNEW.
- Industry/sector labels come mechanically from the Eastmoney quote industry field
f100; do not hand-classify sectors unless the user explicitly asks for a custom theme mapping.
- Count only institution rows:
RECEIVE_OBJECT_TYPE == 001.
- For stock rankings, count unique institution names per stock within the window.
- For sector rankings, aggregate stock-level institution counts to the Eastmoney
f100 industry field, then sort industries by total count.
主要贡献股票 is generated mechanically: within each industry, sort stocks by institution count descending and display the top contributors.
- For weekly sector heat, count unique institution names per stock-week, then aggregate stock-week counts to the Eastmoney
f100 industry field.
- Default cache directory is under the system temp directory. Repeated same-window requests reuse cached raw rows and industry lookups unless
--no-cache is set.
- Keep requests low frequency. The script paginates survey records with a short delay and batches industry lookups; do not loop it repeatedly.
Interpretation
机构数 means unique institutional research participants under the script's de-duplication rule, not the number of separate meetings.
行业 and 主要贡献股票 are data-driven outputs from the script, not model-written summaries. The model may interpret the result afterward, but should not silently relabel industries or contribution stocks.
- Sector heat is an activity measure. The same institution can count once for each stock it researched in the same sector, because the goal is sector-level research intensity.
- Date windows use
RECEIVE_START_DATE, not announcement date.
- Partial weeks can appear at the start/end of a custom window. Call this out when interpreting weekly trends.
- Use
cn-stock-move-reason, stock-sentiment-analysis, or stock-technical-analysis only after a user asks for why a specific stock/sector is hot, whether it is tradable, or whether price confirms the research heat.
Output Guidance
When answering, include:
- Date window and counting method.
- Stock Top 10 with code, name, institution count, industry, and survey dates.
- Industry Top 10 with total heat and leading stocks.
- Weekly industry heat table for the Top sectors.
- A short read of rising/falling heat, while avoiding direct trading instructions.