Implementation of Forecasting Company Project Planning and Realization Using Linear Regression Method

Abstract

Inaccuracies in planning and realizing project budgets are major challenges in construction project management, especially in high market dynamic environments such as PT Dream Island Development (PT.DID) in Canggu, Bali. This research aims to develop a forecasting system that is able to predict project realization value based on planning value and duration, using simple linear regression method. The research was conducted quantitatively with the CRISP-DM approach and analysis of historical project data in 2022-2024. The results show that the regression model produces a coefficient of determination (R²) of 0.9996, which means that 99.96% of the variation in the realized value can be explained by the model. The average realized value of the project was recorded at 79.69% of the planning value, indicating a cost efficiency of about 20.31%. Project duration was also found to be negatively correlated to cost efficiency, with each additional week decreasing the realized value by ±Rp 144,354.88. The project progress model showed an R² of 0.564. Although the model showed high accuracy, there were violations of several classical regression assumptions. This research suggests the development of more adaptive and integrated prediction models, as well as the exploration of alternative prediction methods to improve accuracy in project decision-making.