ROSINA365/-MaizeForecasting

A Shiny web application for forecasting monthly maize prices across five Kenyan counties (Nairobi, Turkana, Marsabit, Mombasa, and Uasin Gishu) using live WFP data. Built to support proactive food security planning using Naive, ETS, ARIMA, and TSLM time series models. # Kenya Maize Price Forecasting Dashboard A Shiny web application for forecasting monthly maize prices across five Kenyan counties to support proactive food security planning. ## Live App **sharonrosina.shinyapps.io ## Overview Maize is Kenya's primary staple food. When prices spike, low-income households are hit hardest. This dashboard provides 6 to 24-month price forecasts powered by live WFP data, giving governments and aid organizations early warning to intervene before a crisis develops. ## Counties Covered County - Zone - Best Model |Nairobi- Urban - TSLM | |Turkana - Arid (ASAL) - TSLM| |Marsabit - Arid (ASAL) - Naive | |Mombasa - Urban - Naive | |Uasin Gishu - Agricultural - Naive | ## Forecasting Models - **Naive** — Uses last month's price to predict forward - **ETS** — Exponential smoothing, weights recent prices more heavily - **ARIMA** — Uses patterns between past and present prices - **TSLM** — Linear trend + seasonal adjustments ## Data Source Live data from the WFP Kenya Food Prices dataset via Humanitarian Data Exchange (HDX). Updated monthly. ## How to Run Locally ```r # Install required packages install.packages(c("shiny", "shinydashboard", "tidyverse", "forecast", "zoo", "lubridate", "plotly", "DT", "httr", "jsonlite")) # Run the app shiny::runApp("app.R") ``` ## Developer - **Name:** Sharon Rosina Wamalwa - **Course:** DSCI 725 — Data Mining for Competitive Advantage - **Instructor:** Thomas Tiahrt, Ph.D. - **Institution:** University of South Dakota - **Year:** 2026