How to compare periods in an analysis
In an analysis you can add the same timeseries twice, over two different periods, and see them laid over each other. That's how you compare operation before and after a measure, or this week against the same week last year.
How to add a comparison
- Open the building you want to work on. See How to switch building and search for a building.
- Open the analysis under Analysis > Timeseries graph.
- Find the timeseries in the list below the graph.
- Click Compare on its row.
- Set Compare from date — the start date of the reference period.
The row now shows Comparison added, and the graph draws both the main period and the reference.
The Reference for comparison column shows which reference the series is set up with.
Note: You only choose the reference's start date. Its length follows the period selected for the analysis, so both curves cover equally long spans.
How to choose the reference date
What you compare against determines what you're actually measuring:
Reference | Answers |
|---|---|
The same period last year | is operation better than last year? |
The week before | did something change recently? |
The date a measure was carried out | did the measure work? |
For that last case: set the analysis period to the time after the measure, and the reference to an equally long span before. The two modes of operation then sit directly on top of each other.
Frequently asked questions
Can I compare two different timeseries instead?
Yes, and that's the common case — add them both to the analysis. Compare is for adding the same series with a time-shifted window. See How to add several timeseries to an analysis.
Why is the reference curve flat or missing?
There's no data in the reference period. Common if you're comparing against a time before the meter was connected.
Can I have comparisons on several series at once?
Yes, but the graph gets crowded quickly. Two series with comparisons make four curves — use Hide on whatever you're not looking at right now.
Does this work in the XY graph too?
The XY graph plots two measurements against each other with no time involved, so time-shifted comparison belongs in the timeseries graph. See How to create an XY graph.
Tip!
Use temperature-corrected series when comparing energy use between two periods here. Without correction you're comparing the weather as much as the operation, and a mild reference period can make a good measure look like nothing at all. See How to use temperature correction and the ET model.