Data requirements
Start with the records you have
Each kind of data answers different questions, and has limits. Whether it supports a valid savings claim is confirmed when we scope the project.
What can my data answer?
| Data available | Questions it may support | Limit to state |
|---|---|---|
| Monthly utility bills | A first whole-building pattern review and, where suitable, a whole-building baseline. | No hourly pattern or equipment split can be read from a monthly total. |
| Whole-site interval readings | Daily and hourly patterns, and out-of-hours reviews. | Whole-site records alone do not identify each equipment load. |
| Equipment meters | Equipment-level patterns and comparisons inside the metered boundary. | Confirm what each meter covers and whether operating context is recorded. |
| BMS and historian exports | Operating and measurement context, where the points suit. | Point names, units, timestamps and meaning need checking. |
| Operating calendar or project record | Defines periods and explains changes. | It is context. It does not replace missing energy measurements. |
How does data become a report?
Existing records
Bills, meters, BMS, files.
Data checks
Interval, units, clock, gaps.
Baseline and analysis
Drivers, fit, comparison.
Report outputs
Results with coverage and method.
How does each source connect?
If your equipment exposes something documented that is not listed, tell us and we will say plainly whether we can read it.
Building management systems
BACnet/IP, BACnet MS/TP and Modbus TCP or RTU, collected by a gateway on site and forwarded outward. Existing trend logs are read as history where the BMS keeps them. Radr reads points; it does not drive them.
Pulse and Modbus meters
Electricity, water, gas and thermal meters. Cumulative registers are handled as cumulative, including rollover and reset.
Historians and data loggers
An existing historian, monitoring system or logger, through its export or API where it has one.
Utility bills and interval data
Half-hourly or quarter-hourly data from the supplier where available, and monthly invoices where not. A meter that cannot be reconciled to its invoice is itself a finding.
Weather
Reanalysis data at the site coordinates, used as a baseline driver.
CSV and file upload
For backfill and retired systems. Timestamps, units and intervals are declared, then checked against the data.
What every feed is checked for
The interval is derived from the timestamps, not the declared setting.
Series with matching shapes are checked for duplicates.
Submeters are summed against the main; the difference is reported as unmetered load, never allocated away.
Units, signs and cumulative registers are checked, and so is the clock.
The result is a written note on each series: its interval, coverage, gaps and what it can be used for.
Which source is best?
A live feed, then a historian export, then a file, then bills: finer grain and lower delay rank higher. Starting from bills is fine; they support a whole-building baseline, not a plant-level answer. Starting without monitoring
Discuss our data
A point list, an export or a set of invoices is enough to start.