Traffic Forecasting using Time-Series Analysis

Traffic jams are a common phenomenon all over the world, especially in a densely populated country like Bangladesh. Due to this, people try heart and soul to tackle this problem by any means necessary to save time to reach their desired destination. Hence the traffic related research is a hot topic now a days which will be quite beneficial for all people living in congested cities. We also tried to do some research on the traffic network to find the most suitable traffic forecasting model to forecast or predict the future traffic value using time-series forecasting models. The only topic which deals with both, traffic prediction and traffic control is traffic time-series analysis for which it is essential. In this paper, we have obtained a suitable dataset containing data of the number of various vehicles for each hour for seven days straight. We have used this dataset to feed into a few time-series forecasting models of our choosing. The models or algorithms considered are ARIMA, ETS, SNAIVE, PROPHET and the last one is the combination of all models we named it "mix". The study shows us the significant difference between each of the models and which one produces a more reliable and accurate prediction.