Rapid and contextual activity analysis : a semi-automated activity category, time, location, and video data collection and analysis methodology

The performance of construction projects is significantly impacted by on-site labor and the productivity thereof. Despite the benefits from technological advancements in recent decades, construction projects are still labor intensive, and labor is one of the most flexible and largest cost factors in a construction project. Thus, a major concern of construction project management has been labor productivity and its improvement. To improve it, labor productivity must be measured and analyzed. One way of doing so is through activity analysis - known as an extension of traditional work sampling. Activity analysis measures the efficiency of the workers' time usage at a construction site. Increasing labor efficiency usually has a positive relationship with higher construction labor productivity. Therefore, activity analysis is considered a major labor performance assessment technique in this research. The objective of this research is to develop a semi-automated data collection and analysis methodology to enable fast and contextual activity analysis. More specifically, this research focuses on the man-machine balanced on-site data collection and the automated data analysis with abundant contextual information to support the interpretation of analysis results for labor productivity improvement study. The prototype of the proposed methodology is implemented and validated with actual datasets from different construction sites. The prototype system proves capable of collecting data efficiently at construction sites and to analyze it in an automatic fashion. This system is shown to provide abundant contextual information related to the activity analysis results. A project manager can quickly and easily find issues related to their high or low labor performance with various scenarios. The indexed videos also successfully provide information about what/how construction workers were performing work at that point. This information can support productivity improvement planning and expedite the continuous evaluation and improvement process of activity analysis to improve labor productivity.