Link to Slides: https://docs.google.com/presentation/d/1oEH9fJnjLwakBKbNXQ9ZD1vRyH4noltmxNf_i_6WVp8/edit?usp=sharing
Please see Issues for discussions and feel free to open a new topic or contribute to an existing one: https://github.com/recurve-methods/comparison_groups/issues
As of 2020-09-25 we have released gridmeter, an open source Python library designed to perform stratified sampling for the purposes of constructing comparison groups. The code is available in this repository in the gridmeter folder; the library's features are documented in an introductory tutorial notebook. Additional features and functionality will be added over time; contributions are welcome. Please raise an Issue with any questions or feedback.
Those of us working in the efficiency and flexibility business have seen COVID-19 upend most of our industry’s assumptions about energy consumption. As an industry rooted in forecasts of avoided energy use (and the ability to measure results on the backside), we now realize that forecasting even basic energy use, not to mention avoided energy use, will be laden with uncertainty for months if not years to come.
With yawning budget deficits ahead of us, the use of public funds for energy efficiency will be closely scrutinized by climate skeptics, agnostics, and provocative gadflies. Private companies willing to make investments will need to know that they are achieving an expected rate of return, especially if interest rates begin to rise and capital becomes more expensive. More than ever, we need sound methods to support claims that energy efficiency is worth doing.
Years ago, Recurve (then Open Energy Efficiency) embarked on a similar journey with the development of the CalTRACK methods. Through two rounds of methods development and consensus building, we helped formulate the specification for savings calculations that now informs program M&V in states across the country. Last year we helped launch the Energy Markets Methods Consortium (EM2) as an independent entity to govern the adoption of revisions to CalTRACK and to support the development of complementary methods in two separate tracks, GRID and SEAT.
These last two months have reinforced -- indeed, provided a clarion call -- for further development of consistent, replicable, and transparent methods to quantify the impacts of energy efficiency programs while also accounting for exogenous factors that fall outside of forecasted normal conditions.
We believe that this work serves the public good. It is not to anyone’s benefit for one entity to come up with a better solution to these heady challenges only to encloak a new set of best practices behind a veil of secrecy. Beginning in two weeks, Recurve will host a twice-monthly open call, where we will describe the testing that we are doing in support of methods to define non-participant comparison groups. In real-time we will be developing a set of replicable, transparent methods for quantifying exogenous factors that lie outside of forecasted consumption trends. We plan to spend approximately 4-6 months doing this work. In the end, we expect to submit a set of methods to the GRID steering committee, and ultimately to the governing board of EM2.
It is our hope that these methods will be available for use in M&V plans for programs planning to launch in 2021. The only thing we know for sure is that patterns of energy consumption will remain highly uncertain for a while. We are confident that the constructive community dialogue that helped inform the CalTRACK methods will also help inform these new GRID methods. We invite you to join us in participating in the bi-weekly calls and bringing your own experience working with comparison groups to our collective efforts, and in conducting your own testing of existing methods for robustness with post-COVID data.
We have selected Friday, May 22nd, from 11:30-12:30 PT as the kickoff event for this working group. Subsequent meetings will be held at the same time on alternating weeks going forward. If you would like to be added to the calendar invitation for the kickoff and subsequent meetings, please fill out the following form:
https://forms.gle/f6iDv9XouNzM5B4F8
On behalf of all of us at Recurve, we would like to thank you in advance for helping us move these critical methods forward during difficult circumstances. We look forward to sharing our work and reengaging with our long-time collaborators.