Engineer the meter panel
Clean half-hourly readings, reconstruct tariff events, and store analysis-ready data in Parquet.
Data Analytics / Causal inference study
Smart-meter data, causal inference, and simulation brought together to evaluate a dynamic electricity tariff rollout.
The context
Using the Low Carbon London trial, this case study evaluates whether a dynamic time-of-use tariff can support a grid investment decision. The work separates measured customer response from rollout reliability and economics.
The approach
Clean half-hourly readings, reconstruct tariff events, and store analysis-ready data in Parquet.
Apply difference-in-differences with fixed effects, placebo tests, and robustness checks.
Simulate virtual substations, household bills, and economic scenarios before preparing a decision memo.
Inside the repository