Small farms are at risk in America. Recent trends show that "more than half of all farmers have lost money every year since since 2013", while large farming companies are able to survive because of their savings. The average family farm doesn't have billions of savings like the large companies, and as such they aren't able to invest in modern technology for their farms.
Farming has always been at the forefront of technology, for it's the innovations in agriculture that have enabled us as a society to grow far beyond the initial villages we originated from. Agriculture has long benefited from technology, and it's not right that family farms can't afford the same technological advantages the large companies can invest in.
A key problem with farming is the sheer overwhelming nature of the job — while anyone can plant seeds and maybe see something grow, it takes a farmer with experience and skill to leverage hundreds of acres to grow and ensure sustainability for their farm. That's why we built Demeter - a tool for simplifying and organizing agriculture to help farmers manage sustainable and efficient farms.
What is Demeter?
Named after the Greek goddess of the harvest, Demeter is a centralized source of truth for farmers to gain insights into their land. Our software allows farmers to optimally plan farm layouts and maximize efficiency and revenue. The user is able to indicate plots of land where they want to plant and indicate the crop they want to plant there. Based on real-time data about air quality, soil and weather conditions, Demeter will determine how suitable a particular area is for planting a particular crop, and provide a general estimation on the expected value of a plot.
1. Find your farm on Demeter!
2. Select the crops you want to plant, and plant them!
3. Foresee future weather, future seasons, and future harvests with our prediction timeline.
4. Click on any plot of land to get soil, fire, humidity, air, and pollen data about it.
5. Check out our complete dashboard to get Demeter's full statistics on your whole farm!
- Realtime, Relevant Data: Demeter provides all the data that you need to make the best decisions for your land - soil moisture, weather, humidity, and more!
- Recommendations Tailoreded to You: Not all land is the same, and it's important to make sure you plant the right crops for the right plot. Demeter's visualization software + agriculture analysis ensures that you'll be able to make the best decisions for your land.
- Anticipate the Future: Use the best in artificial intelligence to see what your farm looks like in the future. -Crop Rotation Advice: Demeter helps you figure out what plants you should grow next, to ensure that your soil stays properly nutriated. -** User-Friendly User Interface*: Demeter was designed with a user-first mindset (using the **Double Diamond* process!) to create an easy-to-use app for all farmers to use!
We utilized the Double Diamond Process; this design method not only includes visual design, but a full-fledged research cycle in which you must discover and define your problem before tackling your solution
Put people first. Start with an understanding of the people using a service, their needs, strengths and aspirations.
Communicate visually and inclusively. Help people gain a shared understanding of the problem and ideas.
Collaborate and co-create. Work together and get inspired by what others are doing.
Iterate, iterate, iterate. Do this to spot errors early, avoid risk and build confidence in your ideas.
UK Design Council
We start off by creating user personas as a method of secondary research. If we had additional time, we would include primary research by heading out to local farms and asking community members about their concerns. By creating user personas, we are able to more deeply relate to different perspectives and understand goals and wants.
|User Persona 1||User Persona 2|
After creating user personas, we move onto visual design - this is where we are able to start formulating our solutions. We utilize the design tool Figma to protoype our designs before doing any coding. Through this, we are able to get iterative feedback so that we spend less time re-writing code. We create Low Fidelity Prototypes first and slowly work ourselves upto High Fidelity prototypes which deeper visuals and interactions.
|Lo-Fidelity to Hi-Fidelity Design|
|Low Fidelity Prototype|
|High Fidelity Prototype|
Demeter is was built with four core components — the frontend, insight api aggregators, the apis themselves, and our custom analysis module.
Our system was thoroughly designed for scalability in mind.
We built a custom analysis module in Typescript to aggregate the research we did. This module is leveraged throughout our application through yarn workspaces, enabling team members to simultaneously work on application development and data science. The frontend then took this data and visualized it using Recharts. This library combines React and D3 to form a reliable, composable, and powerful charting library.
Demeter's complex data visualization toolkit includes multi-line graphs, bar charts, pie charts.
Our backend is built in Node.js using Express, and makes use of two APIs: ambee, which provides information about soil quality, air quality, fire risk, and weather in a given location, and WeatherPlanner, which gives more weather information further into the future. Here is also where the data gets processed using the aforementioned analysis module before being sent to the front-end for presentation.
At the core of the user experience is our web app, built with React.js and deployed with Netlify. We utilized Ant Design as our CSS library and coupled it with SASS (SCSS) as our CSS preprocessor. Our web app allows farmers to optimally plan farm layouts and maximize efficiency and revenue and allows users to indicate plots of land where they want to plant and indicate the crop they want to plant there. This was made possible through a fusion of both Mapbox, WebGL, and ThreeJS, allowing us to overlay the users' potential farms over the very land it would live on.
The tech stack that we used to create Demeter.
Demeter wouldn't be possible without the millennia of human research spent on agriculture. In order to make our predictions and surface the best data for farmers, we needed to gather
- the most common types of crops used in industrial / family farming
- common types of irrigation, filtering by those most affordable
- pricing information for the cost to seed an entire acre for the aforementioned crops
- how to properly factor in soil temperature and moisture levels along with weather data to give appropriate environmental scores. We collected a wide variety of data and learned a lot about agriculture.
We have a full citation of sources here, but here's an aggregate of the various publications we built Demeter off of.
There were many challenges that came with building Demeter: drawing 3D models of plots based on user selection onto a map, processing hundreds of data fields and displaying them on a single page in such a way that is concise and easy to understand, doing the research to understand that data ourselves, etc. We are glad to have been able to conquer these hurdles to build a cohesive product.
What We Learned
We learned about the extensive amount of information that someone needs to know about when farming on an industrial scale, like irrigation techniques, what kind of fertilisers best work with particular plants, the number of different values that are used to quantify how much water, and fertiliser plants use, and so much more. It was genuinely eye-opening to learn that there was so much behind farming, and it definitely made us respect the work behind the agriculture industry a lot more.
A lot of the information that we display on Demeter can definitely be fine-tuned - for instance, the metric that show how much water a crop needs had been simplified far more than it should've been, as a crop's water and nutrient needs change throughout its lifetime, and these are important changes that a farmer definitely needs to account for. Integration with at-home sensors would also be another direction we could've taken, as farmers may want a more reliable source of data on top of what can be provided by the API.
To those trying it - we ran out of free API calls on both ambee and WeatherPlanner and we're too poor to pay for more 😔. We mocked the data, with sample responses we were getting from each API. The user interface is still functioning perfectly but the backend is returning sample data to save us some money. Thank you for your understanding!