From the Client

“The utility industry can face extreme difficulties to model and optimize power production, grid operations, and resource planning.

The vast number of variables and the volume of data required for accurate modeling makes data collection nearly impossible in order to create models accurate enough for enterprise decision-making.

An effective solution is to generate simulated or ‘synthetic’ data rather than attempt to collect actual data. For example, the RUNWITHIT Synthetics team synthesized load curves for our electric customers that closely matched reality during the early months of the COVID pandemic. I know of no other technique that could have predicted such results as quickly with the accuracy we saw.

Synthetic data cuts through complex problems by leveraging the power of statistics. We can quickly generate ‘random’ data that fits the overall behavior of part of a system. We can also quickly generate such synthetic data for each of the many independent parts of a system. These parts can include aspects of customer behavior and social interactions that aren’t easy to measure directly, but can be inferred statistically. By knowing how the parts interact, we can then quickly predict final results by connecting the simulated parts.

The utility industry tends to be conservative with respect to changing how things are done. Basing enterprise decisions on ‘random’ data may go against the grain, but synthetic data and simulation methods deserve the full support of the industry because they are efficient and accurate.”

- Gary Rector

Data Scientist, Information Custodian, SRP, Analytics Center of Excellence