Analytics for Behavioral Insights and Decision Science
The Behavior Analytics team (Pete Cappers, Annika Todd, Anna Spurlock and Ling Jin) in the EIEA Division work in collaboration with the Computational Research Division (CRD) (John Wu and Alex Sim) with a goal to pull customer behavioral insights from granular, high frequency data using a combination of machine learning, behavioral economics and causal inference.
Recent and ongoing collaborations with the Behavior Analytics and CRD team have resulted in several proof-of-concept real world applications highlighting data-driven analytics and ground truth verification through causal inference, including:
- A novel application of machine learning algorithms to derive meaningful household segmentation for time-based rate pilots that relate not only to program enrollment but also actual realized peak period energy reductions;
- A review and quantitative evaluation of machine learning methods applied to extracting residential customer energy attributes;
- A derived library of archetypal 24 hour load patterns from over 30 million residential daily load shapes;
- Technical support to novel program evaluation methods;
- Algorithms to develop and ground-truth detection of household air conditioner ownership and usage.
- Identification and application of innovative analytics and visualization on the survey responses to develop new insights into transportation behavior.
Our work also includes the WholeTraveler Transportation Behavior Study. This study is designed to explore the energy implications of behavioral factors associated with adoption and use of emerging transportation technologies and services (connected and automated vehicles, mobility-on-demand, electric vehicles, e-commerce). The project uses an innovative, regionally-focused survey designed to understand the relationship between pivotal population characteristics, attitudes, and preferences, and their likelihood to adopt emerging technologies and services. In addition, the survey is designed to shape an understanding of how those technologies and services are likely to be used, how these uses are expected to affect the transportation system, and what the resultant energy implications may be.
Our object is to explore the question: How does the US traveler (segmented by demographics) make decisions impacting transportation energy use in the:
- Very short-term: reroute, mode choice
- Short-term: Day-ahead travel planning
- Medium-term: Vehicle ownership and type
- Long-term: Housing location, etc.
We will also identify historic patterns in lifecycle trajectories and map out relationships to transportation behaviors to be used to predict change-points and decision points when people would be most likely to respond to policy incentives.
The work will also couple definitions of heterogeneous traveler groups based on lifecycle trajectories with data on other dimensions of heterogeneity including personality/psychological traits, environmental preferences, metrics of risk aversion and intertemporal discounting, traditional demographic data, and other historic behavior patterns (such as technology adoption) to determine the most useful definition of heterogeneity that can best explain variation in behavioral outcomes of interest: openness to CAV and/or EV adoption/use, car ownership patterns, degree to which TNCs are compliments or substitutes to car ownership or public transportation use, and short-term, high-resolution travel behavior patterns (locational GPS data).
The approach taken in this study involves a survey-based data collection, and subsequent analyses to answer a variety of research questions. The survey is focused in the 9 core counties of the San Francisco Bay Area (Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma). The sampling method used is an Address-Based random sample in this region.
Behavior Analytics: behavioranalytics.lbl.gov
Lazar, Alina, Ling Jin, C. Anna Spurlock, Annika Todd, Kesheng Wu, and Alex Sim. "Data quality challenges with missing values and mixed types in joint sequence analysis." 2017 IEEE International Conference on Big Data (Big Data). Boston, MA, USA: IEEE, 2017.
Kim, Taehoon, Dongeun Lee, Jaesik Choi, C. Anna Spurlock, Alex Sim, Annika Todd, and Kesheng Wu. "Predicting Baseline for Analysis of Electricity Pricing.“ International Journal of Big Data Intelligence (IJBDI) Vol. 5.No. 1/2, 2018 (2017) 3-20.
Jin, Ling, Doris Lee, Alex Sim, Sam Borgeson, Kesheng Wu, C. Anna Spurlock, and Annika Todd. "Comparison of Clustering Techniques for Residential Energy Behavior using Smart Meter Data." AAAI Workshops - Artificial Intelligence for Smart Grids and Buildings, March 2017. San Francisco, CA, 2017.
Cappers, Peter, C. Anna Spurlock, Annika Todd, and Ling Jin. Experiences of Vulnerable Residential Customer Subpopulations with Critical Peak Pricing. 2016. LBNL-1006294.
Jin, Ling, C. Anna Spurlock, Sam Borgeson, Daniel Fredman, Liesel Hans, Siddharth Patel, and Annika Todd. Load Shape Clustering Using Residential Smart Meter Data: a Technical Memorandum. 2016.
Houde, Sebastien, and C. Anna Spurlock. "Minimum Energy Efficiency Standards and Appliances: Old and New Economic Rationales." Economics of Energy & Environmental Policy 5.2 (2016). LBNL-1006327.
Cappers, Peter, C. Anna Spurlock, Annika Todd, Patrick Baylis, Meredith Fowlie, and Catherine Wolfram. Time-of-Use as a Default Rate for Residential Customers: Issues and Insights. 2016. LBNL-1005704.
Spurlock, C. Anna, Peter Cappers, Ling Jin, Annika Todd, and Patrick Baylis. "Go for the Silver? Evidence from field studies quantifying the difference in evaluation results between “gold standard” randomized controlled trial methods versus quasi-experimental methods." 2016 ACEEE Summer Study on Energy Efficiency in Buildings 2016.
Patel, Siddharth, Sam Borgeson, Ram Rajagopal, C. Anna Spurlock, Ling Jin, and Annika Todd. "Time Will Tell: Using Smart Meter Time Series Data to Derive Household Features and Explain Heterogeneity in Pricing Programs." 2016 ACEEE Summer Study on Energy Efficiency in Buildings 2016.