A NOVEL DEMAND FORECASTING METHOD USING LOGISTIC REGRESSION AND CONJOINT ANALYSIS FOR PREDICTING THE VOLUME OF NEW TRANSPORTATION ACTIVITIES

Abstract

In developing new transportation projects, the common approach is to prioritize infrastructure first, with the expectation that demand will follow. This is particularly true for public sector logistics activities, such as seaports, airports, and logistics zones. Unlike the private sector, predicting future demand for public transportation is more challenging due to its complexity. Incorrect demand forecasting can result in improperly sized infrastructure and inefficient human resource allocation. This paper introduces a novel method for forecasting demand in public sector transportation projects. It integrates historical data with the perceptions of transportation stakeholders, improving forecasting accuracy. The method combines conjoint analysis with disaggregated forecasting for each product group. It was applied to a real case namely the Indonesian Ministry of Transportation's plan to develop a cargo transshipment terminal at Denpasar Airport in Bali. A survey of 233 logistics professionals in Jakarta, Bandung, and Denpasar was conducted, and the forecasts were verified through interviews. The results showed that the forecasts for each product category were accurate. This method’s reliability suggests its potential for use in other transportation development projects.