Back to Blog

Creating Data-driven Models for Building Occupancy Estimations

Image of Laura Miller
Laura Miller

In 1792, Robert B. Thomas came up with a secret method for predicting weather throughout the next year and shared his predictions in a publication he called the Old Farmer’s Almanac. Still published to this day, the almanac boasts an 80% accuracy rate and is used by farmers, gardeners, and weather watchers across the United States.

Data-driven models for building occupancy estimation are trying to do the same thing as the Old Farmer’s Almanac: predict and prepare for the future. But smart technologies mean you don’t have to rely on secret methods. The most accurate modeling comes from advanced analytics with machine learning (ML) capabilities integrated into a building management system (BMS). 

Why Occupancy Matters

Demand-based systems show great potential in energy conservation efforts within built environments, especially when demand is based on accurate occupancy data. This is especially true for lighting and HVAC systems, which use most of the energy a building consumes and can easily adjust to variations in occupancy. For example, lighting can dim and HVAC systems can limit functions automatically during periods when few people are in the building, or even in specific rooms.

But energy efficiency isn’t the only benefit of data-driven models for building occupancy estimation. Accurate real-time and predicted occupancy data allow buildings to respond to human needs and help building owners, staff, and tenants use spaces and systems as effectively as possible.

Developing Data-driven Models of Building Occupancy Estimations

Real-time and historical data from a BMS is essential for developing data-driven models for building occupancy estimations. Analytics applications are needed to clarify the relationships between such environmental conditions as indoor air quality (IAQ), temperature, humidity, and how these change as occupancy rises and wanes. However, the volume of data produced by building systems is vast, particularly in a smart building with an extensive network of sensors. A focused approach makes it easier to develop accurate forecasting models. 

In addition to historical and real-time building data, outdoor environmental data is also invaluable. Indoor environments are directly affected by the weather conditions that immediately surround them. As such, wind speed and direction, temperature, and humidity data can all enhance a model’s accuracy. Merging weather and climatic conditions over time with historical building data provides deeper insights into what is happening within a building and what is likely to happen in the future. 

Environmental Sensor Data

Environmental sensors are a cost-effective means of gathering occupancy data. An analytics platform mines this data using ML algorithms to determine the best way to adjust energy usage, temperature, ventilation, and other building variables. The efficacy of such a system is highly reliant on the specific ML algorithms used. 

A 2017 study, for example, found that extracting and selecting the most useful features within sensor data improves occupancy estimates. The researchers discovered that using a ranking-based incremental search within algorithms achieved quicker computations and better efficiency. With this ranking system, they were able to more easily validate the ML algorithms and classify data more accurately. 

Many modern sensing devices connect to controllers via WiFi, allowing them to be placed nearly anywhere. Devices capable of connecting via WiFi are also affordable, in part because they require less infrastructure to install. This makes them an excellent tool for developing custom sensing networks in existing infrastructure. Researchers have found that innovative WiFi-based solutions can outperform existing techniques for sensing occupancy while also being less intrusive, more cost-effective, and more reliable. 

Beyond Modelling: Smart BMS

Smart BMSs integrated with advanced analytics are not only instrumental in creating data-driven models for building occupancy estimations, but also understanding the human dimensions of built environments. 

An intelligent BMS:

safety-c


Enhances comfort, safety, and productivity.

control-d

Expands your ability to manage energy usage and automated systems.

maintenance-c

Improves maintenance.

cost

Reduces operating costs.

With the right technology, you make occupancy models meaningful and create more efficient, responsive, and profitable buildings.

Buildings IOT offers innovative products and services that help you develop data-driven models for building occupancy estimations. To learn more, contact our team of experts today.

 

CONTACT US


Related Posts

Guide to Building Pressure Standards for Hospitals and Medical Facilities

Image of Laura Miller
Laura Miller

A heating, ventilation, and air-conditioning (HVAC) system is one of the most important components...

Read more

Predictive Maintenance vs. Preventive Maintenance: Understanding the Difference

Image of Brian Kolhoff
Brian Kolhoff

Proactive rather than reactive maintenance is quickly becoming the norm for enlightened facility...

Read more