Time series analysis of road accidents in Uganda (2018-2022)

Abstract

The study was about time series analysis of road accidents in Uganda for the period 2018 to 2022 with the main objective of the study being to identify trend patterns of road accidents from 2018 to 2022. The study was based on the annual crime data collected by Uganda Police under the traffic police department that was recorded monthly for the respective five years. Runs test, least squares regression and simple linear regression were used to test for stationarity, seasonality and linear trend respectively. Scatter plots were used to identify the trend patterns of road accidents for the years 2018 to 2022 thus the time series multiplicative model (the product of trend, seasonal and irregular components) was implemented to forecast the monthly number of road accidents for the year 2023. The findings of the study showed the presence of an exponential trend in the observed data of road accidents, there was also a seasonal component in the months of May which was probably due to the prevalence pandemic of covid19 from 2019 to 2020. The study also revealed the absence of stationarity with in the series using the runs test for stationarity. The time series multiplicative model (the product of trend and seasonal component) forecasts show gradual increase and decrease in the monthly number of road accidents with December having the highest number for the year 2023 in Uganda. The study majorly recommends the government to provide and set up more road signs and symbols like zebra crossing, parking, no parking, hill and valley sign posts on road sides which will help to alert road users and control the increasing patterns of road accidents