Skip to main content

Energy Usage Profile Analysis Engine

Algorithm purpose​

This Compute-to-Data algorithm analyzes and compares the energy behavior of supply points grouped by usage typology (e.g., public lighting, residential buildings, garages, public facilities, commercial installations, etc.).

Its main goal is to identify deviations, inefficiencies, and anomalous behaviors within each homogeneous group, establishing clear references that support optimization processes, investment planning, and ROI assessment.

The algorithm does not expose individually identifiable data. Instead, it provides safe statistical comparisons that show how each supply behaves relative to the standard of its profile.

What the algorithm does (high-level overview)​

1. Classifies supply points by usage profile​

The algorithm groups supply points according to user-defined or case-specific criteria, for example:

  • installation type
  • supply function
  • expected operating pattern
  • other relevant contextual variables

Each group represents a comparable energy profile, avoiding comparisons across non-equivalent uses.

2. Computes reference metrics per profile​

For each profile, the algorithm calculates aggregated indicators such as:

  • average consumption
  • typical consumption ranges
  • observed maximum and minimum values
  • the group’s statistical distribution

These metrics define a reference behavior for each usage typology.

3. Measures point-level deviations relative to the profile​

Once the group standard is defined, the algorithm evaluates each supply point by calculating:

  • percentage deviation from the profile average
  • relative position within the group distribution
  • identification of out-of-norm behaviors

This analysis helps detect inefficiencies, sizing issues, or unexpected uses without requiring direct inspection.

4. Produces safe, anonymized comparative outputs​

The output includes:

  • deviation indicators
  • band/range classifications
  • aggregated statistics per profile

It does not expose sensitive absolute values, technical identifiers, or exact locations, making results suitable for sharing across multiple stakeholders.

How this algorithm supports the Compute-to-Data model​

1. Comparison without exposing raw data​

All processing happens in the secure environment. Only relative and aggregated metrics are exported, preserving confidentiality.

2. Establishes standards before optimization​

Before investing in corrective measures or infrastructure, it is essential to answer questions such as:

  • What is considered “normal” consumption for this usage type?
  • Which supplies deviate significantly from the standard?
  • Where are improvements most likely with lower investment?

This algorithm provides objective references.

3. Reduces decision-making errors​

It prevents incorrect comparisons between different uses and reduces the risk of decisions based on intuition or decontextualized data.

Why this algorithm is valuable for the energy sector​

1. Basis for replicable optimization programs​

By defining clear profiles and standards, it supports optimization measures that can be replicated across many similar installations.

2. Direct support for investment and ROI assessment​

Detected deviations help:

  • prioritize actions
  • justify energy investments
  • estimate potential savings
  • evaluate expected economic return

3. Applicable to public and private sector contexts​

It is especially useful for:

  • public administrations (lighting, municipal buildings)
  • infrastructure managers
  • retail and hospitality chains
  • operators of non-intensive networks

4. Preparation for advanced analytics​

Defined profiles serve as a foundation for subsequent algorithms such as:

  • energy optimization
  • advanced anomaly detection
  • scenario simulation
  • strategic planning

Summary​

The Energy Usage Profile Analysis Engine is a Compute-to-Data algorithm that classifies supply points by usage typology and compares their behavior against aggregated group standards. It securely identifies deviations and out-of-norm behaviors, supporting energy optimization and investment decision-making without exposing sensitive information.

Open in MarketplaceDatasetBack to Algorithms