Building Local Farm Alerts with Public Agricultural and Food Data

Agricultural and food data inside government has come out

On August 26, 2026, the Ministry of Agriculture, Food and Rural Affairs, South Korea’s central government ministry for agriculture and food policy, began pilot operation of its Agricultural and Food Big Data Cloud Service (농식품 빅데이터 클라우드 서비스). The pilot runs until October 25, and full operation is planned for late October. This is a roughly two-month stage for checking how people use the service, so some functions may be limited or temporarily unstable.

There are 35 data types that can be connected automatically first. They include wholesale agricultural-product prices, weather observations, local rainfall and agricultural-water reservoir storage rates, materials for specific crop pests and diseases, livestock-disease outbreaks, and eco-friendly certification. This is a specific enough set to test small services that combine prices, weather, and disease information.

The data updates on different schedules. Wholesale prices, local rainfall, and reservoir storage rates update daily; livestock-disease outbreaks and some certification information update in real time; but information on materials for specific pests and diseases updates every six months. Do not group all of this under a single “latest data” label. Show the last update time on each screen instead.

A separate catalogue of descriptions for 166 data types has also been released. It lists each dataset’s content, format, source, and update cycle, but these datasets cannot be connected automatically in the same way as the 35 data types. If the data you need exists only in the catalogue, you must discuss the scope of access with the Ministry of Agriculture, Food and Rural Affairs. For a first service, it is faster to choose from the initial 35.

Individuals can apply as well as businesses and research institutions. After creating an account, applicants must enter the intended access period, purpose and detailed use case, service type such as web or mobile, and the internet address to connect, then receive approval. No stated processing deadline is available, so do not promise a launch date first. Set your schedule only after reviewing the data fields available and the terms of use for each dataset.

Local records become the product, not just the data

Consider Mr. Kim, who runs a farm-supplies shop with three employees in a county-level area. Each morning, farmers call with questions such as, “My pepper leaves are curling,” or “Can I apply treatment before it rains?” Kim has to ask again about the crop, village, recent weather, and materials already used. To find an answer, he switches between weather-alert screens and product manuals.

The call details are scattered across paper notebooks and employees’ phones. When a similar question comes from the same village a few days later, it is hard to find the earlier record. Kim also does not know whether the condition of a farm that bought materials improved unless he calls separately. Important information already passes through the shop, but it is not retained in a form that can inform the next decision.

With automatic data connections, a first screen can show today’s weather alerts, recent rainfall and reservoir storage rates, local livestock-disease outbreak information, and materials for specific pests and diseases together. If the shop also records incoming questions by crop, village, and symptom, “official information” and questions that are actually increasing in the field appear on the same screen. Instead of repeating the same explanations, staff can choose which farms to contact first and which to check again.

But public data alone must not be used to confirm a disease or decide which material to use. Information on materials for specific pests and diseases updates every six months, and crop condition and whether a material can be used must be checked again in the field and against the product label. Confirming photos, talking with farmers, and connecting them to an expert when needed remain human work.

What the service should accumulate is follow-up records: which questions came from which village, what guidance was given, and what happened a few days later. Other businesses can apply for the same public data, but they cannot simply take the local symptom patterns and response outcomes accumulated at Kim’s shop. Continuity of field records is harder to copy than technology.

Relationships remain, too. Farmers already listen to the shops they trade with, and shops can send alerts by text message or messenger to farmers they regularly contact. Rather than trying to gather farmers nationwide from day one, a new service needs to enter these existing contact networks so people actually read it and respond.

The first product does not need to be a huge, all-purpose agricultural information network. Pick one area, one crop, and one repeated question, then put public data and field records on one screen. As farmer responses and action outcomes accumulate, competitors will find it harder to catch up by simply copying the same screen for another area.

Elsewhere, field networks came first

Plantix in India operates a free photo-diagnosis app for farmers. It now aggregates farmers’ diagnostic records by region and provides agricultural-input companies with pest-and-disease trend screens and field-sales alerts, while allowing other services to connect to its diagnostic function.

The company states that the app has more than 40 million downloads and more than 150 million accumulated images, but these are company claims and need independent verification. The important point is not photo recognition alone, but a structure in which cases continue to arrive by region, crop, and season through a free app. Because users and records accumulate together, competitors that build only a similar screen cannot easily reproduce it in a short time.

Esoko in Ghana sent market prices collected by field enumerators, along with weather forecasts and cultivation information, to farmers by text message. Instead of choosing a smartphone app, it used ordinary mobile-phone text messages and a local field-research network, entering a channel farmers already used.

In a study of yam farmers in northern Ghana, selling prices rose by about 5% in the first year. That is roughly the difference between selling goods previously sold for KRW 100 at about KRW 105, but by the second year the effect was no longer clear as the information spread to nearby farmers and traders. Price figures alone become levelled quickly, making this a case for adding trading relationships, field collection networks, and follow-up actions if a service is to last.

Things you can build from this now

1. Ingredient purchasing alerts for side-dish shops

This service shows neighborhood side-dish shops that buy ingredients such as onions and garlic every day both wholesale-price changes and their purchasing records. It is for owners whose ingredient purchases have become uneven after starting delivery sales, and it can begin with a small number of items now that daily wholesale prices can be fetched automatically.

The first screen shows “items up sharply from yesterday,” “planned purchase volume for this week,” and “items to ask suppliers about.” Actual purchase prices accumulated over time and quality records by supplier become more valuable assets than public prices.

2. A pest-and-disease inquiry ledger for farm-supplies shops

This service records crop symptoms reported by phone or in person, the materials suggested, and the date to check back, all in one place. It is for staff at farm-supplies shops in county areas with many pepper and garlic growers, and it is timely to build now because weather and pest- and disease-specific material information can sit beside inquiry records.

The first screen shows “symptoms reported most often today,” “similar inquiries from the same village,” and “farms to ask about outcomes again.” It does not automatically confirm a material recommendation; staff check the product label and field conditions before giving guidance.

3. A shared village irrigation schedule board

This service lets farmers divide water-use dates and share changes while viewing local rainfall, reservoir storage rates, and weather alerts. It is for open-field vegetable farmers who share the same reservoir or groundwater well, and for the person responsible for water management, now that daily water conditions and weather information can be brought onto one screen.

The first screen shows “current reservoir storage rate,” “alerts to monitor ahead,” and “each farm’s water-use order for today and tomorrow.” What is hard to copy is not the weather display, but actual water-use volumes, schedule-change records, and the contact network of participating farms.

4. A shipment commitment board for small producing areas

This service lets farmers post available shipment dates and expected volumes while a coordinator reviews them alongside wholesale-price trends. It is for coordinators at local collection points that combine potatoes or onions from several farms but still coordinate quantities through calls and group chats.

The first screen shows “expected volume by date,” “farmers who have not responded,” and “today’s key wholesale prices.” Anyone can see price information, but records of whether each farm actually keeps its promised volume, and the trust built around those records, remain with the local collection point.

Why this matters where you are

Korea’s government-operated data-access process, including approval requirements and the initial set of 35 automatically connectable data types, is a Korean condition rather than a universal one. Check which local public datasets in your market can actually be connected, how often they update, and what access approval requires. Wherever you operate, local records, existing contact channels, and follow-up outcomes can be more difficult to copy than a public-data dashboard.

Make one call today

Call one nearby farm-supplies shop and ask about three questions that have repeated during the last week, as well as how they record those questions now. If the same question appears at least three times in a week and records are scattered across paper, calls, and messaging apps, that is enough reason to test, for one week, a service that puts local weather and inquiry follow-up on one screen.