Using utility data in RETScreen: getting clean, complete consumption and cost data ready to import

If you run energy analysis in RETScreen, you already know the software is not the slow part. The slow part is the data. Before you can model a retrofit, benchmark a portfolio, or verify savings, you need months of clean consumption and cost figures lined up by meter and by period. In practice that data lives in PDF bills, utility portal exports, and interval files that never quite match. Most teams spend more time keying and reconciling those numbers than they spend on the actual analysis.

This article walks through what RETScreen Expert needs from your utility data, why the quality of that data decides whether your measurement and verification (M&V) results hold up, and how having standardized bill and interval data ready to import removes the bottleneck.

What RETScreen is, and where utility data fits

RETScreen Expert is the clean energy management software published by Natural Resources Canada (NRCan) and maintained by the CanmetENERGY Varennes Research Centre. It covers feasibility analysis for renewable energy, energy efficiency, and cogeneration projects, plus ongoing performance tracking and M&V for existing buildings, factories, and power plants. It is a Canadian government tool, available free in Viewer mode with premium features in a paid Professional subscription.

It is also widely used. As of 2021 RETScreen reported more than 750,000 users in nearly every country and territory of the world, and one independent impact study credited it with more than 8 billion dollars in user cost savings over the years it studied. For all of that reach, every project still starts the same way: with your own metered energy and cost data entered accurately.

The part most energy teams live in is the Performance Analysis Module. It is where you enter measured consumption, add the factors that influence it, build a baseline, and run M&V. NRCan pairs it with a global climate database drawn from about 6,700 ground stations plus NASA satellite data, so the weather side is handled for you. The consumption and cost side is not: that is yours to supply, and it is where the data problems start.

The data RETScreen actually needs

For a performance analysis or M&V study, RETScreen expects three kinds of inputs lined up over the same time periods.

Energy consumption, by meter and period

This is the core series: how much electricity, gas, or other fuel each meter used in each period. Monthly billing reads are the common granularity for facility-level M&V, but interval data (hourly or sub-hourly) lets you normalize more precisely and catch operating patterns that monthly reads hide. The key requirement is consistency: the same meter, the same units, no gaps, and no double-counting when billing periods overlap month boundaries.

Cost data that matches the consumption

Costs turn energy figures into financial results, and they are messier than they look. A single bill blends commodity charges, delivery and distribution charges, fixed fees, riders, taxes, and sometimes credits or true-ups. If you want cost analysis that stands up, the dollars have to be tied to the same meters and periods as the consumption, with the components kept straight rather than lumped into one number keyed from the amount due.

Factors of influence: weather and operations

RETScreen calls the variables that drive consumption factors of influence. Weather is the big one for most buildings, and the module can download weather data and calculate degree days automatically so heating and cooling loads can be normalized. Other factors, such as occupancy, production volume, or operating hours, you supply yourself alongside the energy series.

The bottleneck: hand-keying bills and portal exports

None of the above is conceptually hard. It is just slow and error prone at volume. A single site might have several meters, each with a year or more of history, spread across PDF bills in one format and a portal export in another. Multiply that by a portfolio and the data prep swamps the analysis.

• Transcription errors: a transposed digit in one month quietly skews the whole regression.
• Period mismatches: bills run roughly 30 days but not on calendar boundaries, so monthly buckets need careful alignment.
• Format drift: utilities change bill layouts and portal exports, so a parsing routine that worked last quarter fails this one.
• Interval headaches: Green Button and meter exports come in different structures, with time zone and gap issues that are tedious to clean by hand.
• Reconciliation: making the sum of interval data agree with the billed total for the same period is its own chore.

Why a clean baseline and weather normalization matter for M&V

M&V is where data quality stops being a convenience and starts being the whole game. RETScreen is a common tool for IPMVP Option C, the whole-facility approach, where you measure energy for the entire building before and after a project and attribute the difference, net of other factors, to what you changed.

The method rests on a baseline: a regression that relates your metered consumption to its factors of influence, weather chief among them. In the Performance Analysis Module you enter consumption into the data worksheet, add the influence variables, and RETScreen builds the correlation and reports regression statistics such as R squared and standard error. From there you can weather normalize, run a cumulative sum (CUSUM) analysis to see how performance drifts over time, and produce a chart of baseline versus actual versus savings.

Weather normalization is what makes a fair comparison possible. A mild winter can hide real savings, and a harsh one can erase them on paper, so the baseline has to strip out the effect of weather before you credit or blame the project. That entire chain, from a defensible baseline to a normalized savings number, depends on the input series being clean and complete. Gaps, misaligned periods, and keyed-in errors degrade the regression first, and every downstream figure inherits the damage.

The software is only as trustworthy as the baseline, and the baseline is only as trustworthy as the utility data behind it.

Standardized utility and bill data, ready to import

This is the problem MartinAI is built to remove. Instead of your team keying numbers off PDFs and wrestling portal exports into shape, MartinAI collects utility data from utility connections, Green Button interval files, and scanned or PDF bills, then cleans and standardizes it into one consistent dataset: consumption and cost broken out by meter and by period, in consistent units, with the bill charge components kept separate rather than blended.

The result is data that lines up with what RETScreen and similar analysis tools expect, so the time you spend shifts from data wrangling to analysis. Aligned billing periods, reconciled interval data, and clean cost breakdowns mean the baseline you build rests on numbers you can defend, and a portfolio of sites becomes a matter of loading a prepared dataset rather than a month of transcription.

MartinAI does not replace RETScreen. NRCan's tool remains the place you run feasibility, benchmarking, and M&V. What changes is that the utility data feeding it is already collected, standardized, and import ready, so the analysis starts on a foundation you can trust.