Data Science for the Analysis and Prediction of Energy Consumption in an Industrial Processing Plant

Industrialization remains a key driver of global economic growth, but it also represents one of the greatest pressures on natural resources. The high energy demand of the industrial sector, particularly in processes such as manufacturing, extraction, and transportation, generates significant environmental impacts, including greenhouse gas emissions, water pollution, and biodiversity loss. According to the International Energy Agency, in 2023 the industrial sector accounted for 37% of global final energy consumption, increasing by 4.3% in 2024 with a strong dependence on fossil fuels. In this context, Big Data analysis emerges as a crucial tool in Industry 4.0 to improve transparency and efficiency in decision-making based on big data, characterized by high speed, diversity, and accuracy. Python, with its active community and libraries such as Pandas, is positioned as the preferred platform for handling and analyzing industrial data. Statistical models such as ARIMA and Prophet are evaluated for their predictive capacity in time series; ARIMA excels in linear and stationary processes, while Prophet better adapts to nonlinear trends and multiple seasonalities. Studies suggest that a hybrid combination of both models can optimize predictive accuracy. This article analyzes real data collected using an EMA-Pi data logger in an aquaculture company in Ecuador, applying Python-based tools to identify energy consumption patterns, assess efficiency, and propose energy optimization strategies with an environmentally responsible approach.