Evidence Registry

Academic research backing alternative data alpha claims — cataloguing peer-reviewed papers, out-of-sample validations, and effect sizes across data phenomena.

Evidence Pairs

847

claim-paper links

Papers

312

peer-reviewed

Avg Citations

28.4

per paper

OOS Validated

64%

out-of-sample

Top Phenomena

PhenomenonPapersCitationsAlphaConfidenceSector
1Satellite Foot Traffic
241,840+42bpsHighRetail / Consumer
24 papers1,840 cites+42bpsHigh

Sample Papers

Satellite Imagery and Retail Revenue Forecasting(2021)312 cites
OOS
Geospatial Foot Traffic as a Leading Indicator of Same-Store Sales(2020)287 cites
OOS
Alternative Data and the Cross-Section of Retail Returns(2022)198 cites
IS
2Web Search Trends
312,210+38bpsHighCross-sector
31 papers2,210 cites+38bpsHigh

Sample Papers

Google Trends and Stock Market Volatility(2019)410 cites
OOS
Search Volume as a Predictor of Earnings Surprises(2020)345 cites
OOS
Nowcasting Consumer Demand with Web Search Data(2021)278 cites
IS
3App Download Velocity
18920+55bpsMedium-HighTechnology
18 papers920 cites+55bpsMedium-High

Sample Papers

Mobile App Adoption and Technology Stock Returns(2022)145 cites
OOS
App-Based Signals for Revenue Estimation in SaaS(2021)118 cites
OOS
4Credit Card Transactions
221,650+61bpsHighConsumer / Financial
22 papers1,650 cites+61bpsHigh

Sample Papers

Aggregated Transaction Data and Earnings Prediction(2020)390 cites
OOS
Credit Card Spending as a Macro Indicator(2019)305 cites
OOS
Consumer Spending Patterns and Cross-Sectional Returns(2021)210 cites
IS
5Job Posting Velocity
15780+33bpsMediumCross-sector
15 papers780 cites+33bpsMedium

Sample Papers

Hiring Signals and Firm Growth Expectations(2021)165 cites
OOS
Job Postings as Alternative Data for Revenue Forecasting(2022)120 cites
IS
6Supply Chain Shipping
12540+47bpsMediumIndustrials
12 papers540 cites+47bpsMedium

Sample Papers

AIS Shipping Data and Commodity Price Prediction(2020)132 cites
OOS
Global Trade Flow Signals from Maritime Tracking(2021)98 cites
OOS
7Social Sentiment NLP
281,320+29bpsMediumCross-sector
28 papers1,320 cites+29bpsMedium

Sample Papers

Twitter Sentiment and Intraday Stock Returns(2019)420 cites
OOS
Reddit WallStreetBets and Retail Trading Activity(2022)310 cites
IS
NLP-Based Sentiment Indices for Equity Markets(2020)275 cites
OOS
8Patent Filing Activity
9420+51bpsMediumTechnology / Pharma
9 papers420 cites+51bpsMedium

Sample Papers

Patent Grants and Long-Term Abnormal Returns(2018)185 cites
OOS
Innovation Signals from Patent Filings in Biotech(2021)97 cites
IS
9ESG Satellite Monitoring
11380+26bpsMedium-LowEnergy / Materials
11 papers380 cites+26bpsMedium-Low

Sample Papers

Satellite Detection of Industrial Emissions and ESG Scoring(2022)88 cites
IS
Remote Sensing for Environmental Compliance Monitoring(2021)72 cites
OOS
10Email Receipt Parsing
8290+58bpsMediumE-commerce
8 papers290 cites+58bpsMedium

Sample Papers

E-Receipt Data and Online Retail Revenue Nowcasting(2022)65 cites
OOS
Panel-Based Transaction Data from Consumer Emails(2021)54 cites
IS
11Geolocation Dwell Time
14610+44bpsMedium-HighReal Estate
14 papers610 cites+44bpsMedium-High

Sample Papers

Dwell Time Analytics and Commercial Property Valuation(2021)142 cites
OOS
Foot Traffic Duration as a Signal for Tenant Quality(2022)95 cites
OOS
12Earnings Call NLP
191,100+35bpsHighCross-sector
19 papers1,100 cites+35bpsHigh

Sample Papers

Tone Analysis of Earnings Calls and Post-Announcement Drift(2019)355 cites
OOS
Linguistic Complexity in Conference Calls and Information Asymmetry(2020)280 cites
OOS
CEO Sentiment Extraction via Transformer Models(2022)165 cites
IS

Scoring Methodology

Each evidence pair links a specific alternative data claim to a peer-reviewed academic paper. Alpha represents the median reported excess return in basis points across all studies for that phenomenon. Confidence is a composite score based on paper count, total citations, journal impact factors, and the proportion of studies with positive out-of-sample (OOS) validation. Papers are classified as OOS (out-of-sample) when the study includes a holdout period or independent dataset confirmation, and IS (in-sample) otherwise. All figures are updated quarterly from our literature review pipeline.

Academic Evidence for Alternative Data Alpha

The Vedex Evidence Registry catalogues 847 evidence pairs linking alternative data phenomena to peer-reviewed academic research. Each entry maps a specific data signal — such as satellite foot traffic, credit card transactions, or earnings call NLP — to published papers that quantify its predictive power for financial markets. This registry helps institutional investors and data procurement teams evaluate the empirical foundation behind alternative data products before committing to vendor contracts.

Out-of-Sample Validation and Confidence Scoring

Papers are scored using a composite methodology that weights citation count, journal quality, replication breadth, and out-of-sample validation. Sixty-four percent of catalogued papers include positive OOS results, meaning the reported alpha persists in holdout periods or independent datasets. Confidence tiers range from High (well-replicated, highly cited) to Low (preliminary or single-study evidence), giving procurement teams a quick heuristic for evidentiary strength.