Agriculture

Agricultural businesses face complexity because data is scattered across 3 areas: the farm (IoT + labor), the market (prices + orders), and the supply chain (logistics + quality). PBCP consolidates events from all 3 areas into a single Crop Story and Farm Story.

Business problems solved

Problem	How PBCP solves it
No way to know if this season's plot yield is better or worse than last year	Crop Story → yield history per plot per season
Impossible to monitor hundreds of sensors	IoT event stream → anomaly detection → alert
No idea what today's market price is or whether to sell or hold	Market Intelligence → average price + trend + recommendation
No traceability for where this harvest lot went or whether it passed quality checks	Harvest Story → traceability from field to buyer
Small-scale farmers struggle to access credit	Farm Context → yield track record → credit scoring

Story types

Farm Story

How many plots does this farm have? What was the average yield last year?

Crop Story

When was this crop planted? What fertilizer was used? How much was harvested?

Harvest Story

Who received this lot? At what price? Did it pass quality standards?

IoT Story

Are this plot's sensor readings normal? Were there periods of too little or too much water?

Key events

Event	actor.type	Occurs when
`farm.registered`	`system`	New farm registered in the system
`plot.created`	`farmer`	New plot added with GPS boundary
`crop.planted`	`farmer`	Planting begins; crop type, variety, and area specified
`crop.fertilized`	`farmer`	Fertilizer applied (type + quantity + method)
`crop.irrigated`	`farmer` / `device`	Irrigation performed (manual or automated)
`crop.sprayed`	`farmer`	Pesticide applied (type + quantity + reason)
`crop.disease_detected`	`farmer` / `system`	Disease or pest detected
`crop.disease_treated`	`farmer`	Treatment applied
`sensor.reading`	`device`	IoT reading taken (soil moisture, temp, pH, humidity)
`sensor.alert`	`system`	Sensor value exceeds threshold → alert triggered
`weather.event`	`system`	Heavy rain, drought, storm — from weather API
`crop.harvest_started`	`farmer`	Harvest begins
`crop.harvested`	`farmer`	Harvest complete; yield specified (kg)
`harvest.quality_checked`	`staff` / `buyer`	Quality inspection performed; grade assigned
`harvest.sold`	`farmer`	Sale made; buyer, price, and weight specified
`harvest.delivered`	`driver`	Delivered to buyer / factory / market
`payment.received`	`system`	Payment received from sale
`loan.applied`	`farmer`	Loan application submitted
`loan.approved`	`system`	Loan approved (referencing yield history)
`input.purchased`	`farmer`	Agricultural inputs purchased (seeds, fertilizer, pesticide)
`certification.applied`	`farmer`	GAP / Organic certification application submitted
`certification.approved`	`system`	Certification received

Intelligence use cases

Use case	Signal	Action
Irrigation alert	`sensor.reading` soil_moisture < threshold	LINE notification: "Plot A3 soil is dry — irrigate now"
Harvest timing	Days since `crop.planted` + degree-day model	Recommend optimal harvest window
Disease early warning	High temperature + humidity → high-risk pattern	Alert farmer to inspect the plot
Market price alert	Market price 15% above average	"Today's lime price is good — time to sell?"
Credit scoring	Yield history + on-time payment record	Used by banks to evaluate credit applications
Input optimization	Fertilizer volume vs historical yield	Recommend most efficient dosage
Yield forecast	Weather + historical pattern + growth stage	Estimate yield 2 weeks before harvest

Crop story timeline

crop.planted          → 15 Mar  (Tomatoes, SVR variety, plot B2, 2 rai)
  sensor.reading      → every 1 hr  (soil_moisture, temp, humidity)
  crop.irrigated      → 17 Mar  (auto-irrigate, 450 liters)
  crop.fertilized     → 25 Mar  (15-15-15 formula, 20 kg/rai)
  weather.event       → 2 Apr   (heavy rain 85 mm — risk waterlogging)
  sensor.alert        → 2 Apr   (soil_moisture exceeds threshold)
  crop.disease_detected → 10 Apr (thrips detected)
  crop.disease_treated  → 11 Apr (biological pesticide applied)
  crop.harvest_started  → 20 May
  crop.harvested      → 22 May  (yield: 3,200 kg / 2 rai = 1,600 kg/rai)
  harvest.quality_checked → 22 May (grade A: 80%, grade B: 20%)
  harvest.sold        → 23 May  (buyer: processing plant, price ฿18/kg)
  harvest.delivered   → 24 May
  payment.received    → 25 May  (฿57,600)

Gross margin: ฿57,600 − cost ฿24,000 = ฿33,600 (58%)
Yield vs last season: +12% (from 1,430 → 1,600 kg/rai)

IoT sensor context

PBCP receives sensor.reading events with the following structure:

{
  "eventName": "sensor.reading",
  "actor": { "type": "device", "id": "sensor-B2-001" },
  "payload": {
    "plotId": "plot-B2",
    "readings": {
      "soil_moisture": 42.5,
      "soil_temp": 28.3,
      "air_temp": 33.1,
      "humidity": 68,
      "ph": 6.4,
      "ec": 1.2
    },
    "unit": { "soil_moisture": "%", "temp": "°C", "ph": "pH", "ec": "dS/m" }
  }
}

Alert thresholds are configurable per tenant and per crop type — Intelligence triggers sensor.alert when values exceed or fall below the specified range.

Farm-to-market traceability

Every harvest lot has a harvestId that traces back to the plot → planting date → inputs used → farmer:

harvestId: HRV-20260522-B2
  ↑ plotId: plot-B2 (Farm: Somchai Garden, Nakhon Pathom)
    ↑ crop.planted: 15 Mar
      ↑ Variety: SVR-tomato-001
      ↑ Fertilizer: input.purchased INP-0325
      ↑ Pesticide: biological INP-0411
    harvest.quality_checked: grade A cert
    harvest.sold → ABC Factory, Ratchaburi

Consumers scan the QR code on the packaging → full information displayed.

Full event catalog →

Restaurant index