Even in Canada, there is a huge inefficiency in the production of food. This problem is worsening along with the exponential growth of the human population, and very little seems to be being done about it. We were inspired to improve how farmers collect insights on the current state of their farm by creating a free and easy-to-use open-source tool. We also think that drones and satellites are pretty cool, and wanted to use them in order to hack world issues.

What it does

This toolkit enables users to analyze and view data from satellites and open data sources. The web dashboard allows a user to input latitude, and longitude coordinates to find soil moisture data, vegetation, and soil looseness. We also take data from Quandl to construct visualizations of food prices, farmland expenses, outputs, and productivity over time in the United States. Not only that, but we used regression algorithms to predict the price of various foods for the next five years (such as oranges, soybean, and rice - rice is good, everyone likes rice).

How we built it

AgriGate was mainly built on top of CockroachDB to store satellite and open data from the federal government. During the hackathon, we spent the majority of our time parsing Shape and H5 files before inserting it into the database. In addition, our team pieced together a webUI to display and predict the information collected.

Challenges we ran into

We initially wanted to use drones to survey an area of land to recover photographic data. However, due to complications in software, we were unable to access and parse the data. After the disappointment, our team moved on to complete the rest of our hack by beginning with the Database.

Accomplishments that we're proud of

  • Parsing a ton of data from multiple sources in one interface
  • Finding expansive datasets from obscure locations in the web ## What we learned
  • How to parse Shape and H5 files. ## What's next for AgriGate In the future, our team aims to collect more datasets to improve the accuracy of our Machine Learning Models and increase the resolution of our data.

Built With

Share this project: