Farmer in Uruguay
Climate change is becoming more and more pressing, and is anticipated to reduce crop yields by around 30% across huge swathes of Asia, South America, and Africa by 2050. Population growth is already straining food production, and this will have devastating implications for the future of millions of people living traditional agrarian lifestyles, people often least-equipped to deal with these mounting issues.
Azuremanac is an effort to leverage the power of the Cloud to do everything possible to empower farmers with the knowledge and foresight needed to optimize crop production and practice sustainable farming techniques. Using a text-based service adjusted to account for local language, crop types and sensitivities, and weather patterns, Azuremanac provides the most relevant information and recommendations for farmers. In the backend, Azuremanac is powered by the Azure ML Studio API using deep learning to predict weather patterns for a specific region. This also allows Azuremanac to track the queries sent by farmers and improve upon accuracy and focus in future models.
What it does
The app is a straightforward text-based service designed to be very easy to use. The app ascertains the user's specific regional background and crop focus, then provides personalized recommendations to the user. Since usage patterns are stored in Azure, the app can send alerts to the user in case of insect or crop disease epidemics or natural disasters such as monsoons. It can also recommend the best times to plant specific crops to optimize crop rotation. This can be very powerful for people in rural areas that often lack advance knowledge of these phenomena.
How I built it
The app uses an official UK phone number bought through Twilio. The textual script is powered by the text-based service TextIt. In the backend, TextIt makes API requests to the Azure ML Cloud on which a fully-connected neural network has been trained to predict upcoming crop prices and rainfall. Azure then responds to TextIt, where the app delivers regionalized recommendations for planting, sale, and storage to each user.
Challenges I ran into
I do not have UK phone service, so testing was a little difficult at first. I started the project late, but really enjoyed working on something that I know can greatly benefit the common good.
Accomplishments that I'm proud of
Deploying and integrating new technologies for a cause that can improve the lives of many people and contribute to the greater good. Pivoting an existing project idea that did not work out into something with a lot of potential.