Creates a new calculation job for suggested renovation trajectory calculations.
The Suggested Renovation Trajectory Calculation API enables users to generate a renovation plan for properties, especially when limited details are available. This API enriches the provided data using external sources and predictive models to return a comprehensive renovation trajectory.
The API follows these steps:
Data Enrichment: Enhances incoming property data with additional information from external platforms like Belmap, which provides detailed property measurements using LiDAR datasets.
Prediction Modeling: Utilizes a machine learning model to predict missing property attributes and assess energy performance.
Renovation Trajectory Generation: Generates a suggested renovation plan based on enriched data and model predictions, prioritizing actions that maximize energy efficiency improvements.
The API integrates with Belmap, a platform that provides accurate property measurements using LiDAR datasets. If certain data points are missing (e.g., roof area, building width, depth), the API queries Belmap to retrieve them.
A trained ML model refines missing values and predicts energy efficiency scores. The confidence_score
in the response indicates how much of the final result is based on direct data vs. predictions.
The confidence_score
parameter helps users assess the accuracy of the returned results:
Confidence Score | Expected Price Error (avg) | Standard Deviation | Notes |
---|---|---|---|
< 0.60 | >20% | >30% | High error margin, model relied on many assumptions |
0.60 - 0.70 | ~15% | ~13% | Moderate accuracy, reasonable estimates |
0.70 - 0.80 | ~14% | ~13% | Good accuracy, confidence in most cases |
> 0.80 | ~5% | ~8% | Very high accuracy, minimal deviations |
The suggested renovation trajectory follows a priority-based approach:
For example:
Since calculations are asynchronous, clients should poll the status endpoint periodically until the job status changes to FINISHED
.
confidence_score
. A low value indicates the model had to rely heavily on predictions instead of actual data.Bearer authentication header of the form Bearer <token>
, where <token>
is your auth token.
Body with required input data.
The body is of type object
.
Successfully processed request.
The response is of type object
.