Real-time Occupancy Reporting
Because of the natural error build up that will happen in any occupancy reporting system – no matter how accurate it is – real time reporting should not be treated lightly.
It is recommended that any system providing real-time reporting of occupancy should first go through a period of ‘learning’ of between 2 weeks and 2 months (dependent on the environment and the associated variability of people traffic) where extensive accuracy validations are performed, and end of day correction is employed.
Validations will highlight any individual device issues such as incorrect setup or non-optimal positioning, and as issues are corrected and devices accuracy optimized, the data can then be used to provide live correction factors based on the historical calculated accuracy factors for each week/day/hour.
This can work well in an environment where traffic patterns are relatively similar day by day or on certain days in each week, but accuracy will fall when unexpected traffic patterns are seen, for example around vacation periods, or when there is a public holiday.
In order to correct occupancy figures by factoring, it is necessary to know all the pedestrian counts for the whole day. As above, a day can be considered the period from which the true occupancy is at its minimum to the next period when it is at a minimum - this is usually from midnight to midnight, but might not be in all situations.
In all cases, it is important to set a reset time each day to ensure that cumulative errors do not grow indefinitely. This makes reporting of occupancy in a 24/7 building very difficult.
Some customers are happy to accept real time occupancy figures that have not been corrected and in this circumstance a negative occupancy figure can be replaced by zero. In these cases, a fully ‘corrected’ occupancy report can then be provided at the end of the day.