Stockholm Metro - Energy Monitoring

Presentation:

This project involved the design and deployment of a complete and autonomous energy trend monitoring system (accumulated, generated, and consumed) for the metros of Sweden. The objective was to provide accurate visibility on energy consumption, to predict future budgets, and to automate the generation of essential reports for operational and strategic decision-making. The success of this system led to the regular consultation of Tableau dashboards, reflecting the impact of the innovative scripts deployed.

Achievements:

• Autonomous Project Management and International Coordination: I led this project autonomously, from requirement gathering to design and deployment. This required close collaboration with international teams (Swedish, Indian, and Italian) for requirements integration and coordination.

• Collection and Processing of Energy Data: I gathered and integrated customer requirements for the implementation of a system for ingesting and processing energy signals directly from the trains. I used SQL querying (with complex joins) to extract energy data from the database and Python libraries for their manipulation.

• Design of Prediction Algorithms: I designed and implemented complex Python algorithms for the prediction of monthly and annual energy budgets for each metro line, providing key insights for planning.

• Optimization of Data Storage and Access: I analyzed and compared with Power BI the file formats (CSV, Feather, Parquet, and Pickle) to find an optimal format for energy data. The storage and manipulation of cache files in ElasticSearch were set up, along with an automatic energy cache update system.

• Complete Reporting Automation: I automated the process of generating energy reports (daily, monthly, and annually), including the management of generated reports (copying to a shared folder) and their refreshing (updating in ElasticSearch and Linux/Windows folders).

• Multi-Environment Deployment: I deployed the generation scripts across multiple environments (Dev, Staging, and Production), ensuring the robustness and scalability of the solution.

• Configuration and Customization: I integrated the use of energy configuration files (budget, line names, stations, unique journey references) for maximum flexibility.

• Quality and Documentation: I implemented comprehensive documentation and conducted integration and performance testing to ensure system reliability.

• Workshops and Client Adoption: I organized workshops around the solutions designed with clients and users, facilitating adoption. The generated Tableau dashboards are regularly consulted, proving the success and integration of the solutions within the client's environment.

Technical Stack:

• Languages & Python Libraries: Python (Scikit-learn, Numpy, Pandas, Logger, ElasticSearch, SQLAlchemy)

• Databases / Storage: SQL (Microsoft SQL Server), ElasticSearch (cache files)

• Data Visualization: Tableau Software, Power BI, Plotly, Matplotlib

• File Types: Parquet, JSON, XML, YML

• Version Control: Git

• Development Environments (IDE): PyCharm, VS Code

• Deployment: Dev, Staging, Production environments (Linux, Windows)