Along with the rapid development of today’s social economy, energy shortage has gradually become the focus of national attention. However, most of the construction enterprises lack effective monitoring means, resulting in sloppy energy use. In this paper, based on DBSCAN density clustering algorithm, combined with improved adaptive fast algorithm, scanning data radius parameter to realize polynomial fitting. Cell distance analysis theory is proposed to quickly determine the high-density region. The parameter of contour coefficient is chosen to evaluate the clustering effect of K-center points. Analyze the functional requirements of the energy consumption data analysis platform from four aspects: building energy consumption patterns, discrete cluster points of building energy consumption data, correlation analysis of building energy consumption data, and prediction of building energy consumption data, and design and improve the energy consumption data analysis platform. Comparing the three analysis platforms, the coefficient \(R^2\) of the intelligent building energy consumption data analysis platform designed in this paper is the highest, with an average value of 96.6252%, and the values of Cv are all controlled at 3.00%-7.00%, which meets the requirements of tolerance. Using the platform to analyze the energy consumption data of the actual case, the lighting and socket and power consumption of the test building is relatively stable, and the average monthly lighting and socket consumption is about 112,041,000 kWh, and the lighting and socket and power consumption is regular electricity consumption, which is not affected by the outdoor environment.
