Problem Statement - South Africa in recent years has seen an increase in the price of food. In third world nations such as ours with high unemployment, inequality most of a household’s income is spent on food and shelter. Rising food prices and unemployment will continue to drive our nation's population into poverty and breed other social ills.
The solution - We worked with data obtained from Nemisa which had recorded the monthly prices of wheat, maize and yellow maize since the year 2000 – 2020. Our mission was to visualize the changes in prices over the 20 years as well as build a machine learning model to predict the future prices of these commodities for the next year. We made use of the random forest ensemble machine learning algorithm to ensure the accuracy of our model which recorded a 93% accuracy. We built a web app dashboard on Plotly dash that users can easily make use of to see the time series data and forecasted prices for the next year. We believe we’ve built a tool that decision makers, agricultural economists and FMCG buyers can make use of for their own data driven decision making. We continue to work on improving our model and building a commercial product that business and industry can make use of.
