This is a public repository for our final year capstone project for SE4450 at Western University. The code and more analysis are in private repository. To contact, please reach out to one of our members.
Inspired by the 1 - 100 ENERGY STAR® Score for public and commercial buildings, Residential Energy Performance Rating System (REPR) provides an intuitive and quantitative metric to evaluate the energy efficiency of residential homes.
REPR compares the energy efficiency of the residences and then provides a score between 1 - 100. The higher the score, the home is more energy efficient. Achieving a score of 100 means it is on top of energy efficiency compared to similar homes, and conventionally 75 represents an energy-efficient home. REPR uses multiple statistical methods, comparing energy usages of homes with similar household and geographical characteristics to create a fair score. Thus, the energy efficiency scoring system distinguishes energy-efficient homes normalizing for home-specific characteristics, such as square footage, number of household members, property types, etc.
REPR is accessible through an interactive dashboard and offers distinct features for homeowners and business/energy analysts, all through a secure web application. Once a home is benchmarked, homeowners will have the information to: identify the home’s energy efficiency, improve the home’s energy efficiency, share and report the home’s energy performance. Similarly, the system provides interactive data visualizations to identify patterns and factors for energy efficiency scores for the analysts.
Inspired by Energy Star's Energy Performance Score for public and commercial buildings, our team developed the Residence Energy Performance Scoring, a 1-100 score providing a quantitative comparison between the energy efficiencies of residential homes.
The score aims to provide a benchmark based on similar house attributes and energy usages. The score is calculated by comparing the home's energy usage to its peers with statistical methods to provide a fair score regardless of home-specific factors.
To present use cases of the scoring system, our team developed Residential Energy Performance Rating System (REPR), a software application providing various features based on the scoring system.
To calculate the 1-100 Energy Efficiency Score, we first compute energy usage intensity(EUI). EUI is calculated by dividing the annual energy usage by the total area of the building.
Then, the scoring system estimates the yearly energy usage based on different home factors, regional weather conditions, and metropolis information. These factors influence the energy usage of individual homes but should not reflect on the energy efficiency. For convenience, we refer to these factors as type 1 variables and any other factors that influence the energy efficiency as type 2 variables. Examples of type 1 variables include,
- The population/types of the city
- Types of home
- Square footage and number of rooms
- Years when built
- Weather conditions
The scoring system also uses the type 1 variables to select peers comparison to provide a fair score regardless of home types or geographical location.
The estimated energy usages are used to calculate the energy efficiency ratio with the actual energy usage.
Energy Efficiency Ratio = Actual EUI / Expected EUI
This ratio represents the relative measure of how energy-efficient the home is by comparing the actual energy usage to expected energy usage given different building or geographical factors discussed above(type 1). When the ratio is high, the residence building is less energy efficient, and vice versa.
The energy efficiency ratio is compared with multiple peers with similar home attributes by fitting a probability distribution. Scaling the cumulative probability from the fitted distribution provides a score between 1-100, where 100 represents the home is the most energy-efficient among its peers. The peers are found using statistical clustering analysis with the type 1 variables. To view detail example of how to calculate the score, please see the IPython notebook.
Score = (1 - CDF(Energy Efficiency Ratio)) x 100
Overview of Scoring System
To build tranditional 1-100 ENERGY STAR rating, CBECS(US) and SCIEU(Canada) datasets were used. These data are from government surveys to understand energy usage in commercial and institutional buildings. Similarly, the government also did RECS(US) and SECMURB(Canada) to sample energy usage in residential buildings. We performed analysis on RECS to build our score system
The REPR is a web application developed to demonstrate different use-cases of the Residential Energy Performance Scoring system. The application includes a dashboard for
- Homeowners who want to evaluate the energy efficiency of their home and receive general recommendations to improve the score.
- Analysts want to gain further insights into the driving factors of residential energy efficiency.
The features for typical homeowners who want to know their home's energy efficiency could use the REPR user dashboard to create Residence Portfolio to view annual energy efficiency scores and monthly projected scores.
Furthermore, the user could fill out a quick survey to receive simple recommendations which might help the user improve the energy efficiency of the user's home, which analysts could generate based on the score.If advanced users want to see more detail analytics based on the score, the REPR analytic dashboard demonstrate use case of the score system to analyzes data and different factors (type 2). The analytic dashboard provides following visualizations.
To understand average score over a house attribute, the advanced user could look at the bar chart which group the score average by different home attribute values.
To understand the average score across the geographical region, the advanced user could look at the heatmap which shows average score over different section in a city.
To understand what attributes contributes to energy efficiency, the advanced user could view the piechart which shows percentage different values for homes with score >=75 which are considered energy efficient homes.
To view correlation between a continous variables and the score, the advanced user could look at the sample scatter plot along with the correlation between home attributes and the score.