Project Overview
Background
Tackling food hunger issues is the second goal of the United Nations’ Sustainable Development Goals (SDG). Despite continued efforts in tackling this issue, global food security has been decreasing over the past few years. In 2020, between 720 million and 811 million persons worldwide were suffering from hunger, roughly 161 million more than in 2019. In 2022, affordability decreased sharply due to rapid rises in food costs, decreasing trade freedom and less funding for food safety nets. This has led to a wider difference between the top and bottom-ranked countries in the Global Food Security Index (GFSI) since 2019, highlighting the inequality between countries in food security. Thus, we have opted to solve the problem of food insecurity.
User Persona
Under a limited annual budget, the United Nations (UN) has to decide which countries best require support for food security. However, the food security index is only done once a year and does not consider all countries in the world. Predicting the GFSI by entering a country’s features would be useful to prioritize the fund allocation to countries that need the most help. The UN Security Council may use our model by training data from a range of years, which would yield an equation/model to predict the GFSI of a country.
Problem Statement
To create the problem statement, we first discussed ways to measure the food security of a country. We decided to use the Global Food Security Index (GFSI), a well-rounded index that considers affordability, availability, quality and safety as well as sustainability and adaptation. We then brainstormed various features which would affect the food security of a country: GDP per capita, Education Index, Corruption Perception Index, Percentage of Population Vaccinated, Political Stability, Crop Production Index, Export of Goods, Percentage of Food Produce Exported, Poverty Rate, Road Density and Food Security Index (Target).
After considering our user persona, we arrive at our final problem statement:
How might we create a model to predict a country’s Global Food Security Index (GFSI) by entering a country’s features, thus allowing UN Security Council members to weigh the importance of assisting this country?
Solution
We applied data transformations as well as other data cleaning techniques to build an appropriate Linear Regression model, to predict GSFI based on certain features. Visit links below to view the project with more detailed explanation.
Github
Data Analysis
Frontend Web
