Home/Service Charge Data
Dataset API + Download ML-Powered

Service Charge Data

Real-time service charge intelligence powered by machine learning and building-level aggregation across the UK property market

Coverage
1.2M+
Updated
Real-time
Accuracy
87-99%

Package with other datasets

Combine service charge data with registered leases and ground rent for complete leasehold cost analysis.

Dataset Description

Service Charge Data is a continuously-evolving intelligence layer that provides annual service charge data for leasehold properties across the UK. Built on our property graph, the dataset combines verified records with ML-powered estimates that improve automatically as our network expands.

When available, we surface exact recorded service charges from our property model. For properties without direct records, our machine learning models calculate estimates using real-time building-level aggregation, analysing comparable properties within the same building or spatial proximity. The system self-improves as new property data flows through the network, making estimates more accurate over time.

Methodology

1

Exact records (when available)

Direct service charge records from our property graph. When we have a verified record for a property, we return the exact value with full provenance and confidence metrics.

Verified data95-99% accuracy
2

ML building-level aggregation

Machine learning models analyse real-time property events to calculate estimates from comparable units within the same building. The system processes 2-100+ comparables and learns from each new data point, continuously improving estimate accuracy as the property network expands.

ML-poweredSelf-improving87-95% accuracy
3

Advanced spatial analysis (confidence-gated)

When building-level data is insufficient, ML models analyse nearby comparable properties using embeddings for characteristic alignment. The system considers property attributes, building ownership patterns, and spatial proximity. This method is only activated when confidence thresholds are met, ensuring estimate quality.

Confidence-gatedEmbedding-based80-90% accuracy

Key Features

Real-time updates

Dataset grows automatically as property graph expands with new data events

Self-improving ML

Accuracy increases automatically as network learns from new property events

Building-level intelligence

Analyzes up to 100+ units per building for highly accurate estimates

Confidence scoring

Every record includes accuracy confidence, data source, and calculation method

Usage Example

POST /v1/property/service-charge/estimate
Content-Type: application/json
x-api-key: your_api_key

{
  "locationId": "LDIN-001234567890"
}

Response (28ms):
{
  "success": true,
  "data": {
    "value": 3250,
    "confidence": 94,
    "source": "building_ml",
    "details": {
      "comparablesCount": 18,
      "calculationMethod": "per_sqm",
      "averagePerSqm": 45.2,
      "buildingUnitsAnalysed": 18,
      "accuracy": "87-95%"
    }
  }
}

Data Fields

FieldTypeDescription
valuenumberAnnual service charge (£)
confidencenumberConfidence score (0-100)
sourcestringData source: exact_record | building_ml | spatial_ml
accuracystringExpected accuracy range based on source and confidence
comparablesCountnumberNumber of comparable properties analysed (ML methods)
calculationMethodstringper_sqm | per_bedroom | building_average
averagePerSqmnumberAverage rate per square metre (£/sqm)
comparablesarraySample comparable properties with anonymised details

Use Cases

Conveyancers & Solicitors

Instant service charge intelligence during due diligence and property transactions

Mortgage Lenders

Automated affordability assessments factoring ongoing leasehold costs

Estate Agents & Portals

Enrich listings with accurate service charge data before viewings

Property Managers

Benchmark building charges against market comparables and identify outliers

Investors & Analysts

Calculate accurate net rental yields and total cost of ownership

PropTech Platforms

Build service charge intelligence into search and analysis tools via API

Dataset Info

Type
Verified records + ML estimates
Update frequency
Real-time (event-driven)
Coverage
UK leasehold properties
Formats
JSON (API), CSV (Bulk)
Response time
<50ms (API)

Access Methods

REST API
Real-time lookups via Property API
Bulk download
Full dataset export (CSV format)

Data Quality

Accuracy (exact records)95-99%
Accuracy (building ML)87-95%
Accuracy (spatial ML)80-90%
ML model typeSelf-improving