I recently presented with Bryan Guy of FrontRunner, a software-as-a-service developer, at the Living Future Government Confluence in Seattle. Our interactive presentation focused on the potential of data mining for sustainable buildings and how to undertake a data mining project for your building or portfolio of buildings.
We’ve all heard the adage “you can’t manage what you don’t measure.” However, simply collecting data on building performance isn’t enough – you need to understand it and look for patterns so you can take action.
Data mining is a critical activity to inform a green building strategy—it helps you to understand where your buildings (or occupants) are performing well, and where there may be opportunities for improvement.
What is Data Mining?
There are many definitions of data mining that range from simple to highly technical, so I performed a mini-data mining exercise myself and plugged about 30 definitions into Wordle to generate the diagram above. The key ideas highlighted in the diagram are, “data”, “information”, “patterns”, and “relationships.”
To put it simply, data mining is the process of collecting data from a variety of sources and analyzing it to make better decisions about a strategy.
An excellent, highly visible example of a data-mining project is right here in Seattle. The Fremont Bridge bicycle counter tallies the number of cyclists that cross the bridge each day and pulls weather data from the University of Washington so that city planners can study cycling patterns. The data is used to inform bicycle infrastructure planning.
What to Mine
First, let’s look at the aspects of a building that can provide performance data. We’ve grouped them into two categories – tangible and intangible.
Tangible building features include things such as the HVAC system, lighting, irrigation, water fixtures, and equipment. It also includes areas of the building, such as common areas, restrooms, lobbies and parking lots.
Data you might collect from these tangible features include building-wide energy usage via utility bills or submeters; landscape watering schedules via your facilities manager; or occupancy patterns of common areas and cafeterias via sensors or in-person audits.
Intangible aspects of the building have to do with occupants and include occupant performance, health, happiness, employee attraction and retention and tenant attraction and retention. Viewed over a 30-year period, design and construction and operations and maintenance only account for 8% of the cost of a building. The rest – 92% – relates to salaries, benefits and employee expenses.
According to the Whole Building Design Guide, an additional $2 per square foot per year for brick and mortar costs would pay for itself if it generated a modest 1% increase in productivity. Investing in sustainable features that improve the productivity, health, happiness and retention of occupants can pay off. Data sources for these aspects include in-house surveys, published studies or measurement using samples or estimation methods.
Organizing Your Data
Now that you know what data sources are available in a building, you may be wondering how to organize all this data. You can take a lightweight or heavyweight approach depending on the number of buildings you are managing, your staff capability and resources.
A lightweight approach can be as simple as creating an Excel spreadsheet and manually entering data from your utility company along with information such as ENERGY STAR scores. By reviewing this data over a period of time, you can look for anomalies and investigate whether an issue with a building system or occupant behavior is causing a spike in resource usage. This is an effective approach for smaller, single building projects where you may not have the budget to implement full scale automation systems.
Heavyweight approaches include integrated building intelligence systems that interface with a building automation system and output the data into a custom dashboard. These powerful systems are an option for large buildings or portfolios and can automatically respond to system performance issues.
Charles Kettering stated, “A problem well stated is a problem half solved.” The most effective way to begin your data-mining project is to come up with a goal statement. What green building strategies do you want the data mining process to inform?
At Paladino, we like to use the SMART standard to define project goals: specific, measureable, achievable, relevant and time-bound. An example of a SMART project goal might be cutting back energy usage by 20% in six months or achieving a LEED rating by a certain date.
Relevant building-specific goals might be driven by corporate level sustainability goals for carbon or energy reduction, or third party reporting structures such as GRI, Carbon Disclosure Reporting or LEED.
From there, you can look at the data sources available to you and the methods you need to employ to collect and analyze the data.
For instance, if your goal is to reduce energy, you may need to access utility bills, interview building staff to discuss operations and conduct occupant comfort surveys.
Then, it’s time to develop a game plan—conduct your research, analyze the data against your SMART goals and use it to create a strategy to get your desired result. You should also validate your results at the end to make sure you achieved your goal.
Data-mining can be as simple or complex as you need to provide enough information to make strategic decisions. It’s a broad topic, but I hope this overview is a useful primer on what data-mining is and how you can use it to improve your building performance. Let me know in the comments—how do you think data-mining could help your green building project?
Candice Bullard is Associate Green Building Consultant, FMP, LEED® AP O+M with Paladino and Company in Seattle.