Data Quality Framework

Every data source
earns its place

Truflation indices are only as good as the data behind them. Every provider passes a rigorous two-stage evaluation before a single data point enters any index.

Explore Indexes Become a Provider

Provider Score

AA — Core

out of 100

Accuracy

4.8/5

Coverage

4.2/5

Depth

4.5/5

Delivery

4.0/5

Stage 1 passed · Stage 2 scored

Trusted Data

A powerful competitive edge in the market

100+M

Data Points

100+

Data Sources

Unique Data Access

Over 100+ million data streams are processed in real-time aggregated from more than 100+ different sources.

AI · Data ingestion

ML adaptersSource normalization

Wide Coverage

Truflation boasts the widest coverage in the market, setting it apart from any other data provider.

AI · Index construction

Dynamic weightingAutomatic re-weightingMetadata enrichment

Data Quality

High quality of data sourcing and curation, making it valuable for both off-chain consumers and on-chain data distribution.

AI · Data validation

Outlier detectionManipulation detectionReliability scoring

Growing Ecosystem

Nearly 500,000 monthly visitors, with over 140 developers building on top of the TRUF network.

AI · Distribution

MCP for agentsSDKs & APIsTrading agents

AI Pipeline

From source to signal

Eight automated AI stages carry every data point from raw ingestion to distribution across financial applications and AI agents.

8 stages · fully automated

Data ingestion

ML adapters

Source normalization

Data validation

Outlier detection

Manipulation detection

Reliability scoring

Continuous improvement

Weight refinement

Trust score updates

On-chain integrity

Cryptographic signing

Timestamping

Immutable storage

Index construction

Dynamic weighting

Automatic re-weighting

Metadata enrichment

Forecasting

AI-driven macro prediction

Scenario modeling

Distribution to AI

MCP for agents

SDKs & APIs

Markets & feedback

Prediction market signals

Expectation vs outcome learning

Trading agents

The Truflation Data Provider Credit Score

A structured framework applied to every provider — before they're onboarded and on an ongoing basis thereafter.

Mandatory
evaluation stages

Criteria assessed
before approval

100

Point quality score,
reviewed annually

Start

Provider

A new or existing data provider is considered for use

Stage 1

Credibility

Due Diligence

10 questions. Pass or exit. No exceptions.

Stage 2

Quality

Dataset Scoring

8 dimensions scored 0–100 on a weighted scale.

Outcome

Provider Tier

AA to C rating — reviewed annually

How we source our data

AI · On-chain integrity

Cryptographic signingTimestampingImmutable storage

API Partners

Direct integrations with platforms that expose proprietary data via API.

e.g.ZillowTruliaAmazon

Selected Data Providers

Established research firms and data businesses with verified methodologies.

e.g.NielsenJD PowerIRI

Additional Formats

Structured feeds, scraped sources, and third-party exports.

e.g.CSV feedsWeb dataReports

Stage 1 — Due Diligence

10 credibility criteria

Every provider is assessed against these ten questions using only publicly available information. A single disqualifying finding ends the evaluation immediately.

AI · Data validation

Outlier detectionManipulation detectionReliability scoring

Decision rules

10 / 10Clear pass

6 – 9 / 10Clarify first

≤ 5 / 10Rejected

Legal Identity

A registered business verifiable through a public registry or official database.

Data Origin

Proprietary data collection or a valid license to redistribute.

Methodology

Publicly accessible documentation explaining how the data is produced.

Conflict of Interest

No financial positions that benefit from the provider's own published readings.

Client Credibility

Verifiable use by governments, institutions, or major organisations.

External Citation

Data cited by independent third parties in research or publications.

Leadership

An identifiable, credible team verifiably working at the company.

Business Model

Data provision is the company's core business, not a secondary offering.

Operational Longevity

Actively operating for more than 12 months with ongoing evidence.

Negative Signals

Free from regulatory actions, fraud allegations, or reputational concerns.

Stage 2 — Quality Scoring

8 dimensions of dataset quality

Each dimension is scored 0–5 and multiplied by an assigned weight. Observed Accuracy carries the highest weight — it's the most direct measure of whether the data reflects reality.

AI · Index construction

Dynamic weightingAutomatic re-weightingMetadata enrichment

Data Delivery

Automated and API-based delivery is preferred over manual or report-only methods.

Market Coverage

Broader, more representative coverage of the target market scores higher.

Sample Size

Larger, statistically robust datasets receive higher scores.

Data Completeness

Data must be consistent and complete enough to power a live daily index.

Update Frequency

Daily or intraday updates score highest. Annual or irregular updates score lowest.

Historical Depth

Longer uninterrupted histories score higher. Less than 13 months is disqualifying.

Geographic Granularity

City or postcode-level data scores highest. Global aggregates score lowest.

Observed Accuracy

Highest weight

Correlation with credible independent sources and official benchmarks. The most direct measure of trustworthiness.

Provider tiers

The final score places each provider in a tier. Scores are reviewed annually — providers that fall to Under Review must have an active remediation plan in place.

Core Provider

Score: 80–100

Strategic, deeply integrated sources that form the backbone of Truflation indices.

Qualified Provider

Score: 60–79

Strong, reliable sources with broad coverage and consistent quality.

Conditional Provider

Score: 40–59

Supplementary sources used in specific contexts under ongoing monitoring.

Do Not Use

Score: <40

Provider does not meet minimum quality standards and must not be used in any Truflation index.

Ongoing Monitoring

Scores aren't static

AI continuously refines every active provider's weight and trust score as new signals emerge. A formal portfolio review is conducted annually — and the outputs feed directly into forecasting models and AI agent distribution.

AI capabilities

Weight refinementTrust score updatesAI-driven macro predictionScenario modelingPrediction market signalsExpectation vs outcome learning