How can we base on existing geographic data, open data, and combine with machine learning to predict food security risk in the world? We hope to summarize the various but we believe connected data together to provide a comprehensive analysis of the potential food security risk of a country.

What it does

We build up a model to construct a Food Security Risk Index (FSRI) based on data of population, Climate, Trade/Economics, and Food stress index to provide a possible solution to predict food security, which is completely new and creative idea to present the situation.

How we built it

We select total 11 features from above 4 areas from all countries and construct our risk evaluation model based on Global Hungry Index as our data tag to compute model weightings by linear regression for the model.

Challenges we ran into

Due to the huge amount of data with poor quality of open, we spent a lot of time on data cleaning and data integration. On the other hand, we also encounter technical issue on ArcGIS software.

Accomplishments that we're proud of

A model is constructed and we can play our data on the model. We also can get NDVI from satellite images for agricultural area prediction.

What we learned

The open data sets for food security, the tech. for images processing.

What's next for UnitedSuperCalc

To make our model more robust and enrich the UI.

Share this project: