Fluxora
This research paper presents Fluxora, an explainable and data-driven energy intelligence framework designed for campus buildings. Fluxora integrates data validation, electricity consumption anomaly detection, seven-day short-term forecasting, and deterministic rule-based recommendations into a unified analytical pipeline. The prototype is developed using Python, Streamlit, Pandas, NumPy, Plotly, and SQLite without relying on external AI services. The framework is evaluated using a synthetic one-year campus energy dataset, a manufacturing validation dataset, and 34 educational buildings from the Building Data Genome Project 2 (BDG2). The experimental evaluation examines data validation accuracy, anomaly detection performance, forecasting accuracy, and recommendation reproducibility. Results demonstrate the potential of transparent statistical analytics for campus energy management while identifying limitations related to weekly seasonality and false-positive anomaly detection. The study contributes an openly documented and computationally lightweight framework for explainable campus energy analytics. The estimated energy savings are opportunities and have not been verified through real-world interventions. Keywords: Building Energy Management, Anomaly Detection, Short-Term Load Forecasting, Explainable Analytics, Rule-Based Decision Support, Campus Buildings, Data Validation.
Authors
- Atharva Shelar
- Rahul Gehlot (ORCID: https://orcid.org/0009-0009-4955-4315)
- Atharva Patil (ORCID: https://orcid.org/0009-0008-3129-6963)
- Rajesh Halke
- Sreyash Salunkhe
Institutions
- MIT Art, Design and Technology University (IN)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-15
- DOI
- https://doi.org/10.5281/zenodo.24243529
- Primary Topic
- Building Energy and Comfort Optimization
- Type
- preprint