Network-Charge Forecasting & Monte Carlo

Costing & Risk Intern · E.ON · 2024

What this is

A half-hourly DUoS/TNUoS network-charge forecasting engine improving accuracy ~23% over the prior approach, plus a 10,000-run Monte Carlo for P5/P50/P95 cost outcomes and migration of legacy processes into governed Python, SQL and Snowflake.

Metrics

Accuracy

~23% improvement vs prior approach

Risk

10,000-run Monte Carlo, P5/P50/P95

Automation

Saved ~6 to 7 hours per reporting cycle

Stack

PythonSQLSnowflakeMonte Carlo

What I would not claim

Nothing beyond what is listed above.