Maximizing Facilities Management Efficiency: The Critical Role of Accurate Data in Smart Building Analytics

How often have you heard “We need to do more with less” in your professional career? Probably too many times to count. For facilities management teams, traditionally this meant budget cuts and the need to prioritize critical tasks. However, thanks to technology, doing more with less can now be much more about achieving realistic and increased efficiencies with the assistance of machine learning, artificial intelligence, and data-driven analytics.

As smart building IT and OT systems become more sophisticated and granular with the data they collect, it opens the door to leverage facility data in previously impossible ways. Reaching greater efficiencies in facilities management can be challenging, particularly when accurately interpreting smart building IT and OT data for precise analytics.

Stop Focusing on Analytics

Too often, the emphasis of delivering accurate results through analytics is to focus on the algorithms that process the data. Algorithms are crafted to identify patterns and data and ultimately predict future actions or trends.

For example, an algorithm could be created to analyze weather forecasts and internal/external temperature, humidity, and occupancy sensors that calculate and preemptively regulate an in-building HVAC system. This is a significant advantage, as algorithms are highly adaptable and can analyze various data sources. However, the algorithms and the analytics they produce are only useful when the data being analyzed is clean and accurate.

Good Data = Accurate Analysis

Let’s continue our example of an intelligent HVAC system that adjusts itself based on data-driven insights coming from multiple sources. Our data analysis has operated well for a while, and the HVAC system is accurately adjusting itself to create an in-building environment ideal for occupants.

However, imagine that our IoT sensors, which collect in-building temperature and humidity data, require a newly released software security patch. Once the software patch is applied, the sensors seemingly function normally. But unbeknownst to you, a flaw in the patch accidentally moves a decimal point in the temperature readings, causing bad data to be collected—and ultimately analyzed by our smart HVAC system.

If this goes unchecked, the bad data will soon cause the HVAC system to adjust erroneously, causing in-building temperatures to be too warm or cold. As you can see, these types of intelligent systems and the underlying algorithms are only as good as the data they analyze. Only with good data can you achieve accurate responses.

Baselining, Validation, and Anomaly Detection Are Your Friends

There are three ways to ensure that the data analyzed by your smart building analytics platform is accurate:

1. Baselining. This is the process of tracking data over time, creating a “normal range” where future data is expected. Data deviations that fall outside the range should trigger an alert for further investigation to understand why it’s outside the norm.
2. Validation. This process ensures that data is accurate and meets certain standards prior to being used for data analysis. In certain situations, data can be missing critical information, not properly formatted, or inconsistent when collected from multiple sources. Validation checks scrub the data for errors against a pre-defined set of rules. If those rules are not met, the data is considered “bad” and must be corrected before it can be used for analysis purposes.
3. Anomaly detection. There can be situations where the majority of collected data falls within prescribed baselines and passes validation checks. However, some data sets may contain odd outliers that incorrectly skew analysis results. Anomaly detection can rely on a number of techniques, such as machine learning, statistical methods, or time series analysis, to identify anomalies and remove them from the larger data set prior to analysis.

Good Data Is the Path Toward Data-driven FM

Given all the potential sources of data collection for facilities-related analysis, it is crucial to implement robust data management practices. This is the only way to truly trust the data outputs that can be used to automate facilities’ processes and achieve the desired efficiencies.