Tablón de Goméz is a notoriously coffee-growing area of Colombia, which produces one of the highest quality coffees in the world. In the mountainous terrain, generations of farming families care for the shrubs that produce the coffee seed. It is an artisanal process of harvesting, drying and roasting that brings life, taste and smell to a region heavily affected by civil conflicts.
Even though the coffee is of extreme quality, made by the hands of people with centuries-old knowledge, the children of these farmers will notice that often the pot of coffee is sold for a much lower price than the market. Buyers justify this by saying that they cannot assess the exact amount of moisture or whether the roasting was done properly. Farmers already said that even if they had the tools for measuring, it was not easy to use them and that they did not have time to accompany them with so many tasks in the crop.
Technology and the desire to support their community ‘s work motivated the students of the Institución Educativa Técnica Agroindustrial La Victoria to create CoffeeSmart 4.0, the winner of Solve for Tomorrow Colombia in 2025. The students built an intelligent system with scales, micro:bits (a small programming development board), and sensors that show farmers the ideal drying and roasting levels of the seeds through a smartphone or computer app.
“We saw that the farmers had the tools ready, but because they didn’t know how to use them, they often lost coffee quality,” explains robotics teacher Alexis Rubiel Ardami Ortega. “So, we thought about how to make it easier for our parents to access these tools, helping them make coffee more profitable and receive a fair price.”
The school already has a strong connection to coffee, and the beans are often the subject of laboratory experiments and other STEM-based learning (Science, Technology, Engineering, and Math). By participating in a government program, they won five micro:bits, small, simple computers that gave the students and the mediating teacher an opportunity to design a project that unites technology and the passion for coffee production.
Automating drying and roasting
Even though they come from coffee-farming families, the young people in different high school grades (tenth, eleventh, and twelfth grades, aged between 15 and 18) did not know much about the specifics of coffee production. Fortunately, their research field is just next door, and the coffee plantations are practically laboratories.”
They started with a bibliographic research, to understand more about what determines the ideal quality of coffee, as interviewed by their relatives, who told them about the difficulty of proving that coffee has drying and roasting – these are the two most important processes of coffee – right for buyers.
This is due to a mix of factors: the cultivation processes are artisanal, and even if the coffee is good, the farmers do not necessarily know the correct degree of humidity or roasting, relying on their senses and wisdom for that. Most farmers have the tools to measure both processes, but report that they are not practical.
Take coffee drying, for example: what determines a good drying is the weight of the coffee when it is free from a certain amount of moisture. The tools available to accompany a process that lasts under the sun for 15 or 20 days are neither practical nor compatible with the other tasks of cultivation.
With the systemized complaints, the group understood that their task was to simplify access and knowledge of drying and roasting data in an application, so that farmers know each stage and can be charged for fair payments. For that, micro:bits, coupled with weight and temperature sensors, could provide very simple information to both processes. The automation of tools then began, creating a system.
The balance with micro:bit and a weight sensor is the key piece for the drying stage: “We put our large batch of coffee on a patio and next to it we place our balance. On the scale will be a sample of 200 grams of parchment coffee. We will monitor the weight, which will be in the same condition as the batch. As the coffee loses moisture, it’s losing weight. When that 200 g sample reaches 104 or 105 g, it indicates that the coffee is at the optimal point of humidity, 10 to 12%, and this value influences the final price. So, the idea is that the application can tell on the balance what the correct humidity is to reach the buyers,” said Ortega.
Torrefaction is an equally important process, and the students, who have a coffee lab at school, already had a prior study to understand the torrefaction curve , explains the teacher: “We have a small toaster where samples are tested on the well-known toasting curve. It takes time to take the temperatures to look at the coffee phase, what state it is in. We connected a microbit to a sensor so that it records the temperature and additionally the time, indicating to us the phase of toasting.”
The information from the coffee curve is stored in a database, which shows the organoleptic characteristics of the coffee and ensures that the next roasting rounds keep the record.
The balance and the ideal app
An application connected to the sensors was the means chosen by the students to provide the ideal parameters of the coffee. They had some basic programming experience, but not enough to create an application. They then activated the knowledge of the school principal, who is a systems engineer, who helped them develop a simple platform that can be operated from anywhere in the crop.
To complete the design, the students chose to draw inspiration from the very design of materials used by their fellow farmers: a grated tray securing coffee for final weighing.
Fine-tuning the project with farmers
The final prototype then consists of micro:bits, temperature and weight sensors moved to solar panels, the scale and also any machinery that the coffee maker uses to do the roasting, such as a toaster. With the system soon, it was time to test their product with who can de facto say whether it worked or not: The system was taken for coffee producers.
“They liked it a lot, because they said it was very practical. We explained to them about the app, which is easy to use, ready to download, and sometimes a bit tricky when connecting because it’s in a rural area, but they found everything simple”, explains the teacher.
Many of the people who work with coffee are elderly, so they will ask for the reading instruments in the application to be modified to have larger letters, and so that the information does not come only in text. but also with graphics that show the temperature, humidity and the torrefaction curve, for example. These are small modifications, but they make a big difference and are part of Project Based Learning (PBL).
Even with a rich prototype and a good return, the students still had to prepare a lot to make the stages of presentation of the idea for the program Solve for Tomorrow. Ortega recalls that the most impressive change was the investigative quality of the work of the young people. They had to do a lot of research on coffee and in this process, they saw themselves as researchers, as people who master a subject and know how to talk about it not only for the coffee farmers but also for specialists and buyers.
The future of the Coffee Smart 4.0 system looks good. The prototype became a social innovation, a treasure of the school community together with the crops that are part of the territory, and the idea is that the school can give continuity to the project, making it bigger and reaching other municipalities that work with the same product.
The future of young people also seems long, as Ortega ends: “The teenagers were motivated by what has to do with STEM areas, I looked at them very motivated on that issue, they like technology a lot and innovate every day. If suddenly someone is interested already in medicine, they also look at engineering, as something to be able to achieve in the future, as engineers, programming and all that. It’s really nice to see those teenagers who are getting more motivated in these areas every day.”
Learn more about the project in the following video:
