Farm AI
To develop a platform that uses machine learning to provide foresight into proper and timely farming through predictive analysis. Through our application, farmers will have valuable information regarding the average demand for different crops in a particular time period, the expected average supply of crops, a predictive analysis of the price variations of those crops in a given time period, what crops are potentially most profitable for farmers to grow in a given time period, and more.
There are many countries in the world whose economy is predominantly based around agriculture. Agriculture plays an important role in the economies of a large number of developing countries in Asia, Africa, and South America. Worldwide, the agricultural sector accounts for a significant percentage of many countries' economies. In most parts of the world, farmers grow crops based on their current knowledge and past experiences.
For example, most farmers grow a limited number of crops. This is because they have the necessary knowledge to grow those crops and the experience gained in previous years. If we consider a particular period in which the price of a crop, such as carrots, skyrockets suddenly due to certain reasons, most farmers will opt to grow carrots during the next season because there was a high demand and high price for it previously. By growing carrots, they expect to obtain high profits since they have experience of the price of carrots going up due to a lack of supply to satisfy the demand. In reality, many farmers may think the same way and grow carrots, ultimately resulting in an excessive supply that exceeds the market demand. When there is a high supply of a particular product, the result is usually a drop in price. Therefore, carrots might become less profitable and create losses for farmers who grew them. The excess supply can also lead to overproduction and crop wastage.
The solution we developed to address this problem is to create a platform that uses machine learning to provide predictive analytics for crops. Through this application, users will be able to gain valuable information regarding the expected supply of different crops in a particular time period, the expected demand for those crops, a predictive analysis of the average price variations in the market, and what crops are potentially most profitable to grow for a given period. Users will also be able to gain insights into which crops are best to grow at a given time based on weather conditions, geographical information, and more, thereby reducing overproduction, crop wastage, and financial losses for farmers.
At present, around 40% of crops produced are wasted throughout the world. This is an alarming percentage, especially when we consider the ever-increasing demand for food. As the global population keeps increasing, crop wastage must be controlled. Even in Sri Lanka, the situation is similar, with around 30-40% of crops being wasted due to low yield levels and a mismatch between production and market demand.
We conducted a survey involving 138 farmers and 12 agricultural officers from different areas in Sri Lanka. A major concern of the farmers was that the prices they receive for their crops were low due to overproduction; often, at certain times, either a particular crop gets overproduced or, whenever a crop is not produced enough, a sudden and large deficiency occurs in the market and prices skyrocket. If we consider the overproduction issue, the supply exceeds the actual market demand and consequently the prices the farmers receive are low, and most crops are thrown away and never consumed. On the other hand, when a particular crop is not produced enough, market prices spike and a deficiency occurs, destabilizing the market.
Most farmers are caught in between this disparity and mostly end up not receiving reasonable prices for their crops, leading to wastage. Due to these issues, farming is no longer considered a profitable venture and fewer people are willing to become involved in it. This was one of the major concerns for regional agricultural officers and related government officials, as a large percentage of the children of present farmers choose not to pursue farming because the profitability of it is decreasing. This makes the issue even more serious because even though fewer people pursue farming, the demand for food is increasing rapidly.
Farming is crucially important to the population of the entire world, and we saw that many of the prevalent problems in farming were due to a lack of proper and timely foresight into demand, market supply, and what crops to grow to achieve profits. Therefore, if the much-needed foresight is provided, more people would be encouraged to invest in farming and present farming communities would be able to obtain better profits for what they grow, uplifting their lives as well.
Scope
The project will consist of a smartphone application that provides farmers with important foresight into potential market variations and valuable information on how to pursue farming effectively and maximize profits.
Through the application, farmers will be able to access critical information, including the average expected market demand for a particular crop in a given period, the average expected market supply for that crop in a given period, the expected price variations for that crop, and foresight into which crops would potentially be most profitable for a particular farmer to grow in a specific period.
Apart from these core features, the application will also include additional features such as a message board to notify the farming community about diseases that could potentially destroy their crops, suggestions for mitigating those diseases, and a chat platform that facilitates knowledge sharing among the farming community. In the future, this application could also serve as a platform for buying and selling crops.
Feasibility
Economic Feasibility
There are 570 million farms around the world. If our app provides foresight, the lives of farmers could improve significantly. Therefore, we expect high adoption rates for the app. For example, if we can reach just 1% of farms worldwide, that would encompass 5.7 million farms.
We aim to provide the core forecasting capability of the app free of charge initially. For add-on features, including disease alerts, auctioning, and access to message boards, we aim to get a monthly subscription fee of $20. Given the revenue and scale, we will be able to handle employment costs, maintenance costs, and server costs.
Technical Feasibility
Microsoft ML Studio will be used for the core component of this application. It has neural network components, and we will use recurrent neural networks to develop the application. Microsoft ML Studio and its technologies are subscription-based and already existing technologies. The frontend will be developed using the Cordova and Ionic frameworks. These are open-source, cross-platform supported technologies.
Operational Feasibility
Farmers are the main stakeholders in this platform. In the current world, more and more smartphones are being used globally. According to www.statistica.com, the number of smartphone users by 2019 is expected to reach around 2.5 billion worldwide, which is approximately 36% of the global population and growing rapidly each year. Smartphones are penetrating many developing countries throughout the world. Therefore, this application has huge potential in the market because there are no competitors offering predictive analytics and a range of other features as this app does.