keywords: Fuel (PMS) Price, Forecasting, Machine Learning, Time-series Analysis, Visualization
Prediction in time-series data deals with predicting future trends and movements of data from the previous analysis of the data. Analysis of past events or tends usually influences future events or trends. This is achieved through the use of effective visualization tools/methods. Fuel (PMS) pump price is a time-series data that changes due to government policies and programmes, economic factors and market dynamics. To address this challenge, a visualization based approach is adopted in this work to forecast fuel (PMS) prices for selected states in Nigeria: Kogi, Lagos, Bauchi, Borno, Anambra, Rivers and Abuja - FCT. The dataset used is obtained from monthly fuel (PMS) pump price gathered and monitored by Petroleum Products Pricing Regulatory Agency (PPPRA) and National Bureau of Statistics (NBS) over the period of six years (that is, January 2016 – December, 2021). The result of the time-series graphs showed that there are continuous fuel (PMS) price fluctuation across geo-political regions of Nigeria especially North East region. This is influence by large disparity and consistency in the fuel (PMS) pump price, product availability and enforcement of regulatory policies and programmes during the period of study.