What is this? Heat maps can be a highly effective means of visualizing electric meter consumption data. With each column representing a day and each cell representing 15 minutes, we can clearly see the operating hours of this facility.
Data Engineering
REST APIs
Adam designed a standard that service APIs should be developed to. This included but was not limited to authentication, authorization, and application layout. His primary day to day activity is developing new high performance REST APIs across multiple services to this standard.
Primary Tools Used: Node, Express, Azure App Services
Fault Detection Analytics
Enabling efficient analysis for tens of thousands of pieces of equipment, Adam is the primary developer and maintainer of a fault detection analytics platform specifically designed for building automation system data. The analytics can be viewed and processed in near-real time or scaled down to a lower resolution depending on project needs.
Primary Tools Used: Python, Azure Functions
Utility Data Collection
Through screen scraping, PDF parsing, and tabular imports, Adam has saved hundreds of employee hours by automating utility data collection from 10 different utility providers and counting.
Primary Tools Used: Python
Asynchronous Task Management
In order to perform over 100,000 simultaneous analytics on time-series data, Adam aided in implementation and maintains an asynchronous task management system in Python. This system has the capability of processing data in near real time and has helped in the identification of over one million dollars of savings opportunities to date.
Primary Tool Used: Python, Azure Functions
Classification
Adam has continued to develop and maintain various tools for classifying and managing years of timeseries data for over millions of data points totaling over 75 billion unique records.
Primary Tools Used: Python, TensorFlow, JavaScript, React, Azure Functions
Database Management and Design
Holding data used by many tools, both internal and external, Adam is the primary administrator and designer of a 200 table MariaDB relational database in addition to a dedicated timeseries ClickHouse data store. Types of data stored include building automation system trends, utility bills, asset information, user management, and materialized statistic views.
Primary Tools Used: MariaDB (MySQL), ClickHouse, Azure
CI/CD
Utilizing Azure DevOps, Adam has increased efficiency by automating builds and deployments of both internal and external applications used every day.
Primary Tool Used: Azure DevOps
OpenADR
Referencing the OpenADR 2.0b specification Adam designed and helped implement an OpenADR interface. This interface is utilized to receive demand response events and convert those into operations that get disaggregated and dispatched to buildings.
Primary Tools Used: Node, Express, Azure App Services
Optimization
Adam developed a regression based algorithm and visualizations to inform both chiller staging and optimal ice burn for a large university. The process ran in real time based on current conditions and predicted and optimized the next 24 hours of events. This provided equipment operators with the tools they needed in order to make more data driven decision on how best to operate a 45 thousand ton chilled water and resulted in thousands of dollars of avoided electric demand charges.
Primary Tools Used: Python, MySQL, eDNA, DGLux