DiLytics Powered a U.S. based Agency's Efforts to Improve Energy Efficiency and Compliance with the Implementation of Snowflake Cortex AI

The Overview

The client is a U.S.-based regional energy network agency (Agency) formed through a collaboration of nine counties in the San Francisco Bay Area. It runs several regional-scale energy efficiency programs.

One of these programs empowers single-family homeowners by providing insights into their current energy consumption and guiding them on how to make cost-effective improvements through energy-efficient home upgrades.

The core components of the program are:

Home Energy Assessments: The core of the program involves conducting professional home energy assessments.

Home Energy Score (HES): A key component is the utilization of the U.S. Department of Energy’s (DOE) Home Energy Score. A certified assessor evaluates a home’s energy efficiency and provides a score on a scale of 1 to 10, along with estimated energy consumption and utility costs.

Education: The program educates homeowners about their property’s specific energy usage patterns and potential areas for improvement.

Energy Consumption Recommendations: Homeowners receive customized recommendations for energy efficiency upgrades that can help save energy, reduce utility bills, and improve comfort.

Rebates: The program often provides rebates or incentives to help offset the cost of obtaining a Home Energy Score assessment.

Real Estate Focus: A significant aspect is working with real estate professionals (realtors, appraisers, lenders) to increase their understanding and valuation of energy-efficient homes, thereby integrating energy performance into the real estate market.

Electrification Checklist: The program has also incorporated an Electrification Checklist alongside the HES to identify a home’s potential for switching from gas to electric appliances, supporting decarbonization goals.

By providing clear, standardized information about home energy performance, the program aims to drive energy efficiency investments and contribute to regional energy savings and climate goals.

The Challenge 

While the program was effective in providing homeowners with valuable energy performance information and recommendations through Home Energy Scores (HES) and assessments, managing and leveraging the resulting data presented significant challenges such as:

Fragmented and Siloed Data: Information generated by the program such as assessment findings, HES scores, specific recommendations, rebate applications for assessments, and interactions with real estate professionals often resided in separate databases or systems. This fragmentation made it difficult to create a single, unified view of a property’s or homeowner’s engagement with the program.

Difficulty Connecting Actions to Outcomes: A major challenge was tracking whether the assessments and recommendations led to energy efficiency upgrades. Data on subsequent actions, such as applying for rebates for recommended upgrades, or even real-world changes in energy consumption, were disconnected from the initial assessment data. This made it hard to measure the program’s effectiveness in driving actual energy-saving behaviors and quantifying its real-world impact.

Complex and Time-Consuming Analysis: Answering nuanced business questions that required combining different data points – for instance, analyzing the correlation between specific home characteristics, HES scores, recommendation types, and subsequent upgrade uptake across different geographies or homeowner demographics – required complex data extraction and manual analysis by skilled data personnel. This process was slow and couldn’t provide timely insights.

Limited Data Accessibility for Decision-Makers: Program managers, outreach teams, and partner organizations who were not data analysts found it difficult to directly access and explore the program data. They relied on static reports that could not be easily customized or drilled down into, hindering their ability to get quick answers to ad-hoc questions or investigate specific trends as they emerged.

Obstacles to Program Optimization: Without an easy way to analyze integrated data, it was challenging to identify which aspects of the program were the most effective. This included understanding which types of recommendations were most often implemented, which outreach methods resonated best with specific homeowners or real estate professional segments, or where geographical targeting could be improved. This limited the ability to continuously refine and optimize program strategies for maximum impact.

These challenges created barriers for the Agency, limiting its ability to fully asses the program and make data-driven decisions to improve its reach and effectiveness.

DiLytics Solution – Snowflake, Data Modeling & Cortex AI

To transform the Agency’s data challenges into opportunities for deeper insights and program optimization, DiLytics designed and implemented a comprehensive data and AI solution. The approach involved establishing a robust data foundation in the Snowflake Data Cloud, integrating data from multiple data sources using Snowpark, organizing the data through expert modeling, and providing seamless access via interactive data visualizations in Tableau and a natural language conversational interface powered by Snowflake Cortex AI.

Recognizing the issue of fragmented and siloed data, DiLytics’ data engineers focused on creating a single source of truth, collecting and integrating data from the program, including assessment details, HES scores, recommendations, and associated activities, along with relevant datasets from other programs and flat files.

All this integrated data was then loaded into the Snowflake Data Cloud, which provided the secure, scalable, and centralized platform necessary to consolidate the Agency’s diverse program data.

Within Snowflake, DiLytics further developed a sophisticated relational data model, structured specifically for analytical efficiency, a process critical in defining clear relationships between key entities like properties, homeowners, assessments, recommendations, completed upgrades, and program participation events, thereby laying the groundwork to enable complex analysis and directly addressing the difficulty in connecting actions to outcomes at the data infrastructure level.

With the data unified, cleaned, and expertly modeled in Snowflake, DiLytics then enabled the dissemination of insights via data visualizations in Tableau and a natural language interface (agent) using Snowflake Cortex AI.

This integration of Snowflake Cortex AI, which comes directly with the Snowflake Data Cloud, provided a cutting-edge natural language interface that allowed users to ask questions, like the ones below, in simple English:

The Outcome

The combination of integrated, well-modeled data in Snowflake, the intuitive access provided by Cortex AI, and DiLytics’ expertise enabled Agency to achieve key outcomes such as: