industry-prosperity-tracker

板块景气度查询

What this Skill does

When a user wants to track an industry's prosperity / business cycle status, this Skill:

1. Identifies the industry (built-in sector or user-specified custom sector)
2. Selects representative leading stocks for that industry (from built-in config or AI knowledge)
3. Fetches the latest macro + stock financial data via scripts/fetch_indicators.py (using free public data sources only via AKShare)
4. Calculates a composite prosperity score (0-100) via scripts/calculate_score.py
5. Generates a one-page prosperity dashboard via scripts/generate_report.py using assets/report_template.html
6. Outputs: overall score, direction judgment, indicator detail table, key signals, compliance disclaimer

Architecture: Universal Engine + Dynamic Stock Selection

This Skill uses a generic indicator framework that works for any industry:

Universal Macro Layer (40% weight, same for all industries):

• Leading: Official PMI (15%), Caixin PMI (15%), PPI YoY (10%)
• Coincident: Export YoY (15%), Import YoY (10%)

Dynamic Stock Layer (60% weight, per-industry):

• Coincident: Leading stocks' quarterly revenue QoQ (15%, auto-split among stocks)
• Lagging: Leading stocks' quarterly gross margin QoQ (20%, auto-split among stocks)

Weights are automatically distributed based on the number of stocks provided. No manual weight configuration needed.

How to use

Step 1: Identify the industry

When a user says something like "看医药行业景气度" or "帮我跟踪新能源板块":

1. Check if the industry is in BUILTIN_SECTORS (see scripts/fetch_indicators.py):

   • semiconductor (半导体): 北方华创, 韦尔股份
   • pharma (医药): 恒瑞医药, 药明康德, 片仔癀
   • new_energy (新能源): 宁德时代, 比亚迪
   • consumer (消费品): 贵州茅台, 五粮液

2. If built-in: use directly.

3. If NOT built-in (e.g. user says "看军工行业"):

   • You (the AI) select 2-5 representative leading stocks for that industry based on your knowledge
   • Example for 军工: 600760(中航沈飞), 002179(中航光电), 000768(中航西飞)
   • Pass them via --stocks parameter: --stocks "600760:中航沈飞,002179:中航光电,000768:中航西飞" --sector-name "军工"

Step 2: Fetch data

Run scripts/fetch_indicators.py:

# Built-in industry
python scripts/fetch_indicators.py --sector semiconductor

# Custom industry (AI picks stocks)
python scripts/fetch_indicators.py --sector military --sector-name "军工" \
    --stocks "600760:中航沈飞,002179:中航光电,000768:中航西飞"

# List all built-in sectors
python scripts/fetch_indicators.py --list-sectors

Output: data/{sector}_latest.json

Step 3: Calculate score

python scripts/calculate_score.py --input data/{sector}_latest.json --output data/{sector}_scored.json

Step 4: Generate report

python scripts/generate_report.py --input data/{sector}_scored.json --output output/{sector}_report.html

Step 5: Present

Present the generated HTML report to the user with a summary of the score and key signals.

AI Stock Selection Guidelines

When selecting representative stocks for a new industry, choose stocks that are:

1. Market leaders in their sub-sector (top revenue or market cap)
2. Representative of the industry's value chain (not just one sub-segment)
3. Listed on A-share exchanges (6-digit codes: 6xxxxx SH or 0/3xxxxx SZ)
4. Have at least 4 quarters of financial data (needed for QoQ calculation)
5. Pure-play or industry-dominant (avoid conglomerates where the industry is a minor segment)

Good examples:

• 军工: 中航沈飞(600760), 中航光电(002179), 中航西飞(000768) — covers aircraft + components
• 化工: 万华化学(600309), 华鲁恒升(600426) — covers PU + coal chemical
• 房地产: 保利发展(600048), 招商蛇口(001979) — covers top developers
• 银行: 招商银行(600036), 兴业银行(601166) — covers retail + interbank

Built-in industries

To add a new built-in industry: add an entry to BUILTIN_SECTORS dict in scripts/fetch_indicators.py.

Data sources

All data sources are from the compliance whitelist. See references/data_sources.md for details.

Key principle: this Skill uses free public data exclusively via AKShare open-source library. Data sources include National Bureau of Statistics (PMI, PPI), Customs General Administration (exports, imports), and exchange-disclosed financial reports (stock Financial Abstract). No paid terminal data (Wind, Bloomberg, iFinD) is used.

Compliance

This Skill provides data aggregation and indicator calculation only. It does NOT:

• Recommend buying or selling any security
• Predict price movements or returns
• Provide target prices or investment ratings
• Guarantee any investment outcome

Every output includes the disclaimer: "本工具仅提供数据整理和指标计算服务，不构成任何投资建议。历史数据不代表未来表现。数据可能存在延迟。"

See the bottom of assets/report_template.html for the fixed disclaimer footer.

Scoring methodology summary

• Each indicator is scored: +1 (improving), 0 (flat), -1 (deteriorating) based on MoM/QoQ change
• Weighted sum produces a raw score in [-1, +1]
• Converted to 0-100 scale: score = (raw + 1) * 50
• Interpretation: >60 = upcycle, 40-60 = neutral/turning, <40 = downcycle
• 3-period moving average applied for trend smoothing

See references/scoring_methodology.md for full details.

Output format

The report includes:

• Industry name and reporting period
• Composite prosperity score (0-100) with visual gauge
• Direction indicator (upcycle / downcycle / turning point)
• Indicator detail table (name | latest value | MoM change | direction | weight | contribution)
• Top 3 positive signals and top 3 negative signals
• 6-month historical score trend
• Compliance disclaimer footer