In Colombia’s high-altitude greenhouses where flowers and crops require careful cultivation, farmers are turning to data to stay ahead. With shifting climate patterns and increasing pressures on agriculture, the digital transformation of greenhouses is becoming essential for improving yields and protecting crops.
Leading this transformation is Gabriel Coch, a veteran in mapping and geospatial technologies, with AgroPatterns — a data-driven platform that’s improving how farmers monitor crop health, mitigate pests, and optimize productivity.
Gabriel Coch, Founder of AgroPatterns
Gabriel’s journey into agritech follows decades of experience in digital mapping, location-sharing applications, work in emergency management, and even collaborations with industry pioneers like Ray Ozzie, the creator of Lotus notes. But his work with AgroPatterns has become his most personal and impactful endeavor.
In the cutthroat world of fresh flowers — a $30 billion industry — timing is everything. Growers must anticipate market peaks for events like Valentine’s Day and Mother’s Day, ensuring their products bloom precisely when demand surges. To meet those demands, farmers have long relied on intuition, and local and historical knowledge.
But as climate change introduces new unpredictability, traditional cultivation methods can benefit from real-time data and predictive analytics. Temperature fluctuations, pest infestations, and disease outbreaks can jeopardize entire harvests, and past monitoring methods (which still often rely on paper logs and Excel spreadsheets for many farmers), can be inefficient and slow to reveal critical insights. In an industry where a single pest outbreak can wipe out millions of dollars in inventory within 48 hours, waiting for data to be compiled and analyzed can be a very costly delay.
AgroPatterns gets farmers deeper insights about their crops by integrating traditional manual record keeping with real-time sensor data, predictive analytics, and digital mapping. The core of the platform relies on a structured data collection method designed specifically for greenhouse environments, where every plating row and section is coded for consistency and traceability.
A network of field scouts traverse greenhouses weekly, manually identifying pests and diseases. Instead of scribbling notes on paper, scouts input observations into a simple offline-compatible app, syncing data instantly to the cloud when the device reconnects to signal to transmit its offline data. Each greenhouse is divided into structured zones, assigned unique identifiers rather than relying on GPS coordinates (which drain mobile batteries too quickly in the field). This approach ensures that every data point is linked to a physical location within the greenhouse, creating a reliable dataset that informs decisions.
“We are able to identify problems as quickly as possible,” Gabriel explains. “Catching mites in the egg stage is exponentially cheaper and less toxic than dealing with an outbreak of adults. The goal is to reduce blanket spraying and instead limit spraying to targeted zones, which saves money and reduces chemical exposure for workers and consumers alike.” Observable plays a key role in processing and visualizing the collected data. Gabriel has built a suite of Observable Notebooks, leveraging D3 and Observable Plot for geospatial mapping and using Observable Framework to transform raw field data into actionable visual insights.
An Observable Notebook that Gabriel created to procedurally generate greenhouse maps.
Using geospatial overlays built with Observable and Mapbox, farmers can view real-time maps that display the status of pest infestations, chemical treatments, and productivity metrics at a glance. These overlays replace the previous digitization process, where each field’s layout of latitude and longitude points had to be painstakingly mapped and updated separately. Now, Gabriel’s Observable Notebook automates the entire process — inputting just a few key reference points generates a complete geospatial representation of the greenhouse down to the individual planting beds.
AgroPatterns uses AI-driven automation to make data more accessible for farmers. Its WhatsApp integration delivers real-time alerts and insights directly to field workers, ensuring quick decision making without reliance on in-office dashboards. Thousands of messages are exchanged daily, making this chat-based system the primary interface for field scouts.
AgroPatterns' WhatsApp integration keeps farmers informed on the go.
Workers can log updates using voice commands with AI-powered transcription, eliminating the need for spreadsheets or handwritten notes that can be lost or misrecorded. Farmers can also request AI-generated insights — such as identifying pest hotspots — and receive an instant geospatial map highlighting trouble areas.
“People in the field aren’t sitting in front of dashboards all day,” Gabriel notes. “They need immediate, actionable information.”
AgroPatterns' growing degree days dashboard, built with Observable Framework, predicts crop growth for optimal harvest timing.
One of AgroPatterns’ most powerful tools is its use of growing degree days — a metric that tracks accumulated heat over time to predict plant growth stages. By continuously collecting temperature data via custom-built sensors, AgroPatterns provides farmers with precise insights on when their crops will be ready for market.
“Temperature determines how fast plants grow,” Coch says. “If you’re two days off in predicting flower maturity, that’s the difference between a perfect Valentine’s bouquet and a wasted shipment.”
This level of precision allows farmers to fine-tune their harvesting schedules, ensuring that every stem is cut at peak readiness. Combined with AI-assisted analysis and WhatsApp alerts, the system helps growers anticipate issues before they become costly crises.
AgroPatterns currently serves over 60 farms, covering approximately 15% of the Colombian market. But Gabriel sees a much broader application for this technology. While his current focus is on flowers, he believes the same principles can be applied to other high-value perishable crops.
“This is about improving agricultural practices,” he asserts. “Many small farmers lack technical expertise, but with simple tools, we can help them make better decisions, reduce chemical use, and increase their profitability.”
As AgroPatterns continues to evolve, Gabriel remains a staunch advocate for Observable’s role in his workflow. His ability to iterate quickly, prototype maps, and deploy solutions with minimal overhead has been a game-changer. But beyond technology, what drives him most is the human impact.
“Constant exposure to chemicals has many detrimental health effects on those that work in the field. If we can reduce unnecessary spraying, we’re not just saving money — we’re saving lives.”
For Gabriel Coch, this is more than a business. It’s a mission to make agriculture smarter, safer, and more sustainable — one data point at a time.