> Source: https://ashareapi.com/en/docs/guides/sector-rotation/  ·  Markdown version for LLMs / AI agents

Tutorial

# How to Read A-Share Sector Rotation
 Sectors reveal direction earlier than single stocks: rankings find strength, valuation percentiles judge price, supply chains show the logic. Three endpoints make one sector check-up — code below.

## Why sectors before stocks
 A single stock can move on a headline; **an entire sector moving usually means money is choosing a direction**. Fix the direction first, then pick names inside it — that order works better than the reverse.
 Three steps: **① ranking (who is rising) → ② valuation percentile (expensive?) → ③ supply chain (why)**.

## Step 1: today’s sector gainers

 Python (key required)
```
import requests

KEY = "ct-your-key"
r = requests.get("https://api.ashareapi.com/v1/sector",
 headers={"Authorization": "Bearer " + KEY}, timeout=15)
b = r.json()
rows = b.get("structured") or b["data"]

for row in rows[:10]:
 print(row) # sector name / change % / leading stock / advancers & decliners
```

-
 Covers **industry / concept / region** sectors

-
 Read **advancers vs decliners** first: a sector up with only 1–2 names rising is leader-driven, not broad

## Step 2: is this sector expensive? (valuation percentile)

 Python
```
# Sector codes look like pt01801780 (Shenwan sector)
r = requests.get("https://api.ashareapi.com/v1/sector-valuation",
 headers={"Authorization": "Bearer " + KEY},
 params={"code": "pt01801780"}, timeout=15)
print(r.json().get("structured") or r.json()["data"])
# Returns: PE / PB / PS / PCF + dividend yield + historical percentile
```

-
 **The percentile is the point**: whether PE 20x is high depends on its own history

-
 ≥ 80% → expensive (worse odds); ≤ 20% → cheap (but ask *why* it is cheap)

-
 **Sector codes start with `pt`** (e.g. `pt01801780`), not stock codes

## Step 3: where is the logic? (supply chain)

 Python
```
# Where a stock sits in its supply chain
r = requests.get("https://api.ashareapi.com/v1/industry-chain",
 headers={"Authorization": "Bearer " + KEY},
 params={"mode": "stock", "code": "sh600667"}, timeout=15)
print(r.json().get("structured") or r.json()["data"])
# Returns: chain theme / node / upstream-midstream-downstream / relevance / business description

# Reverse: the chain graph of a theme (which companies, at which positions)
r = requests.get("https://api.ashareapi.com/v1/industry-chain",
 headers={"Authorization": "Bearer " + KEY},
 params={"mode": "graph", "topic": "semiconductor"}, timeout=15)
```

-
 `mode=list` → all themes (**183**) · `mode=graph&topic=X` → graph · `mode=stock&code=X` → single stock

-
 **Why it matters**: when a sector runs, knowing who supplies whom beats reading concept tags

## Combining: "strong + not expensive + has logic"
 |
| | Ranking | Percentile | Supply chain | How to read

| | Strong | Low (≤30%) | Clear | **Best case**: possibly early

| | Strong | High (≥80%) | Clear | Trend may continue but **odds worsen** — trade it, do not invest in it

| | Strong | Low | Vague | **Careful**: possibly pure money game with no fundamental support

| | Weak | Low | Clear | **Watchlist**: logic present, position low, waiting for a catalyst

 Strong
 Low (≤30%)
 Clear
 **Best case**: possibly early

 Strong
 High (≥80%)
 Clear
 Trend may continue but **odds worsen** — trade it, do not invest in it

 Strong
 Low
 Vague
 **Careful**: possibly pure money game with no fundamental support

 Weak
 Low
 Clear
 **Watchlist**: logic present, position low, waiting for a catalyst

## Common mistakes and boundaries

-
 **Wrong sector code** → must start with `pt` (e.g. `pt01801780`); stock codes will not work

-
 **One day only** → single-day sector moves are noisy; look at 5/20-day too

-
 **Treating cheap as a buy reason** → cheap can be a value trap (a declining industry)

-
 **Relevance is not a position** → high relevance means business overlap, not price synchronicity

-
 **Rate limits** → sector endpoints need a key; anonymous access covers only the five free endpoints

## FAQ
 How do I get sector codes?
 They look like `pt01801780` (`pt` + 8 digits). Use `/v1/sector` for the list of sector names/codes, or `/v1/industry-chain?mode=list` for supply-chain themes.

 How is the valuation percentile computed?
 It compares the current PE/PB/PS against that sector’s **own historical series**, giving the current value’s percentile (80% = more expensive than 80% of its history). **It describes position, it does not predict direction.**

 How is a supply chain different from a concept sector?
 A **concept sector** is a market label (some members merely ride the concept). A **supply chain** is organised by real upstream/midstream/downstream relations with a **relevance score**. For logic, the chain is more reliable.

 The sector is up but my stock is not. Why?
 Three common reasons: ① the move is leader-driven and you do not hold the leader; ② your stock is concept-related, not business-related (check chain relevance); ③ money is rotating inside the sector (high to low).

 Are these endpoints free?
 **No** — `/v1/sector`, `/v1/sector-valuation` and `/v1/industry-chain` all require an API key. The free five are quote, kline, hot, market overview and breadth.

 Last updated: 2026-09-23
 Code and fields come from live responses (verified 2026-09-23: sector returns 124 rows; sector-valuation accepts pt codes; industry-chain stock mode returns theme/node/position/relevance). Sector code format follows the endpoint docs.

 Read next

-
[Endpoint reference (incl. sector / sector-valuation / industry-chain)](/en/endpoints)

-
[Quantifying A-Share market mood (all-free endpoints)](/en/docs/guides/market-mood)

-
[Python top-trader board data: 4 approaches compared](/en/docs/guides/longhubang-data)

-
[Error codes & rate limits](/en/docs/errors)

 [← All tutorials](/en/docs/guides)