Smallholder farmers have the potential of producing a myriad of data but it’s mostly undocumented. Take an example of a piece of information that goes uncaptured; what acreage of land they’re farming?
Only 10 % of Africa’s rural land is registered, that leaves a whopping 90 % that’s undocumented and informally managed; farms included. Many farmers do not have access to measuring tools, so they would be missing the exact size of the land they farm.
But that’s just the beginning.
As subsistence farmers, they do not treat farming as a business. They grow food crops mainly to feed their families, therefore, crop output is quite low and only a little surplus is available for trade within local markets.
Also, there is no tendency to keep track of their farming operations, letting information such as details on inputs, fertilizers, and pesticides slide into the ether — uncaptured. The same thing applies to how much crop they produce and the volume that gets wasted along the agricultural value chain due to insufficient storage and transport capacity.
While organizations are trying to gather data, data still exist in silos. Some silos are opaque, while others are striving for openness.
Take, for example, Consultative Group to Assist the Poor (CGAP) that drew from a reservoir of nationally representative household surveys across six countries and provides a dashboard with access to more than 300,000 data points on smallholder families’ financial lives. The main goal of this would be to help financial service providers to learn about the demand for financial services through smallholder communities.
Another example is aWhere that provides a rich set of agronomical insight layers on top of weather data to know how weather affects crop growth and health.
Imagine building a shareable data ecosystem that combines not only those two datasets but others including the market value of crops, and demand for mechanical services. That would allow for more successful deployments of more innovative products.
The Importance of Contextualization
Another problem that smallholder farmers face is that even when they do exist, data dashboards are designed with product and service providers in mind, without localizing and contextualizing for first-mile farmers. Also, it’s usually not accessible in a format farmers would understand unfortunately due to their low literacy levels.
Before we address how Big Data can be utilized to advance African farms, we need to address the basics, starting with a means to digitize farming operations that start with planning, goes through growing, post-harvest handling and linkage to markets. By capturing data points that happen at each stage of the farming process, we start to build intelligence on what exactly happens at the farm level.
As reported by CTA, potential benefits of collecting data for smallholder farmers include increased participation and self-empowerment, improved or new products such as logistical, extension, financial, input and trade services, and higher crop yields. However, for farmers to be able to make informed data-driven decisions, they need answers to very specific questions such as, “What does the market want now? What should I farm? What’s the cost of inputs? How’s the weather going to affect my farming operations?”
So whatever tool that needs to be built should put the small holder farmer at the center of its design, be in a format they can access, and a language they understand. It should also be contextualized in such a way that they are able to practically apply the knowledge on their farms immediately.
Understanding the need for farmer-centric design, Kuza built its entire methodology and technology on field-based market intelligence.
Take, for instance, Kuza’s Agri-entrepreneurs dashboard, and how, in addition to giving access to over 5,000 HD micro-learning video modules, it tracks and digitizes over 26 interactions between them and the smallholder farmer. Cased within Kuza’s IoT Backpack, the AE carries it to the farmers thus allowing them to capture and record specific questions at the farm-level.
As each AE manages approximately 200 farmers and there many AE’s impacting over 100,000 Indian farmers, that’s a lot of data being generated from 7 Indian states.
Data-Driven Future of Farming
While the first step of making the best use of data is through digitizing and tracking farm operations, the next level of data falls under machine learning to predict — crop production, market prices, and crop demand.
For that, data collection can be done through the installation of IoT devices on the farm to optimize and manage the water and fertilizers used; a process known as precision farming.
By using modern technology such as sensors, GPS-based soil sampling, and drones, farming can be controlled and made more accurate. The benefit of this is managing resources such as water and quality inputs, and controlling pests and diseases from spreading enough to ruin farms. The Centre for Agriculture and Bioscience International estimates that about 49% of Africa’s crops are lost to pest and diseases each year. By collecting images of potential insect pests, and training an artificial intelligence model, apps can be used to identify potential threats and send alerts. This would help avoid massive crop failures.
What is Hindering Data Sharing
Despite the huge potential for Agro-related data to deliver value to smallholder farmers, many problems hinder data sharing and the creation of a shareable data ecosystem. For example, data needs to be interoperable, thus requiring a common vocabulary for the collection and utilization of data within relational databases. For example, CGIAR uses a thesaurus called AGROVOC, which was developed by the UN Food and Agriculture Organization, and is currently aligned to 18 open datasets related to agriculture.
Also, multiple tensions exist that hinder the effective usage of data such as competing commercial interests, data ethics, user privacy, and data security issues. Different stakeholders will need to overcome those issues together and learn from each other about best practices, before onboarding their data on a single ecosystem.
Ultimately, building a data ecosystem will give stakeholders the power to predict and produce more food more sustainably. This will be done by empowering analysts to mine information for trends and allow for robust, accurate recommendations for farmers, thus eventually leading to global food security.