Basic terminology for anyone who wants to consume Spacewell’s occupancy data.
- Occupancy data classes
- Occupancy time slots
- Time slots sizing
- Monitored spaces
- Space classes
- Space capacities
Occupancy data classes
Various brands and models of sensors can be used to track occupancy in office buildings. These devices are categorized as “presence” or “headcount” sensors.
Presence sensors
These are devices that track if a space is occupied or not by detecting motion. Therefore they are also called “motion sensors”.
The output is either 1 (presence detected) or 0 (no presence detected) on a timeline.
Headcount sensors
These are devices that track how many persons are occupying a space.
The output is a positive integer value on a timeline.
Occupancy Data Class
As a result, Spacewell Data Platform groups the recorded occupancy data in 2 Occupancy Data Classes:
- PRESENCE
- HEADCOUNT
These values are built-in and cannot be changed.
Occupancy time slots
All occupancy sensors register data on a timeline.
However, the sampling frequencies are different for each sensor brand, type and model. Some sensors register presence multiple times per minute, others only every 10 minutes, some even irregularly in function of motion detection.
Spacewell Data Platform records all these raw sensor data without exception. Reporting on these is possible by using a combination of these views:
- sensor_data_v1
- sensor_devices_v1
- locations_v1
Reporting on raw sensor data is very technical and contains outputs at irregular intervals depending on sensor brand, model, type and configuration. Reporting and analysis on raw sensor data is challenging and requires data science.
For most customers, Spacewell Data Platform curates the raw sensor data and exposes them in so-called “timeslots”.
These are data values on a timeline per monitored location, which everyone understands:
- Per 15 minutes
- Per hour
- Per day
All the occupancy data tables contain a primary key “timeslot_id”, which is a concatenation of location_id + calendar_datetime. This combination is always unique.
When using time slots, you take away the learning curve of dealing with raw sensor data.
Time slots sizing
As a data engineer, depending on your business needs, you have access to the 15-minute, hourly and daily data sets.
How many rows are in the 15-minute data set?
Each year has 365 days x 24 hours x 4 quarter-hours. That is 35040 rows per year. For 1 sensor.
If your organization has 1000 sensors, that is a whopping 35 million rows per year!
Data analysis doesn’t always require 15-minute details. At least not for a long period of time.
Spacewell Data Platform also aggregates and exposes occupancy data per hour and per day.
Per hour
Each year has 365 days x 24 hours. That is 8760 rows per year. For 1 sensor.
If your organization has 1000 sensors, that is still 8.7 million rows per year!
Per day
Each year has 365 days. That is 365 rows per year. For 1 sensor.
For an organization with 1000 sensors, this leads to 365.000 rows per year.
Depending on the number of spaces with sensors installed, consider these figures:
Legend:
Monitored spaces
The occupancy timeslot tables return details per monitored space, not per sensor, for the following reasons:
- A sensor is not always linked 1-on-1 to a space
- In some use cases, 1 headcount sensor can monitor multiple spaces
- In other use cases, a set of 4 sensors attached to individual seats may lead to a calculated headcount of the room they are part of
- Spacewell can only create timeslots for spaces that are monitored. In reality, a floor or building may contain more spaces than monitored.
- etc
Therefore, all the exposed occupancy data are per space, not per sensor.
Metrics like occupancy, utilization, efficiency and others are always calculated based on the monitored spaces, not on the total spaces in a building.
Space classes
In Spacewell Studio, customers can map their operational “room categories” to a predefined set of “Space Categories”, for example:
In the exposed data set, Spacewell has relabeled these Space Categories to “Space classes”:
Space Class | Meaning | Studio name |
|---|---|---|
WORK | Spaces where people can sit to work. Can be closed personal offices or seats in a landscape office. | Work spaces |
MEETING | Enclosed rooms for meeting / collaboration purposes. May have a nr of seats but these cannot be booked individually. | Collaboration spaces |
FOCUS | Enclosed rooms for focus work, typically silent areas | Concentration spaces |
SANITARY | Sanitary spaces like individual toilets. | Sanitary spaces (*) |
PARKING | Individual parking spots for vehicles, inside a building or on a terrain. | Parking spaces (*) |
OTHER | Anything else | Unclassified spaces |
The timeslot data sets contain a column “space_class” for analysis purposes.
It is essential to understand that the values are time-dependent!
The timeslots contain the values that were actual at the time of recording.
If a space belongs to class “WORK” on 01/01/2024 and changes are made to the Space Class mapping on 01/10/2024, which may lead to another class “MEETING”:
- Between 01/01/2024 and 30/09/2024 the Space Class will be “WORK”
- From 01/10/2024 the Space Class is “MEETING”
(*) Space Classes “SANITARY” and “PARKING” are also available in the timeslots, but should be excluded for office occupancy analysis.
Space capacities
The timeslot data sets contain a column “Capacity” for analysis purposes.
Capacity represents the number of seats inside a space.
In Spacewell Studio, customers can enter their operational “room capacities”, for example:
It is essential to understand that the values are time-dependent!
The timeslots contain the values that were actual at the time of recording.
If a space has capacity 4 on 01/01/2024 and someone later decides to change this to 6 on 01/10/2024:
- Between 01/01/2024 and 30/09/2024 the Capacity is 4
- From 01/10/2024 the Capacity is 6