Skip to content
Alternative Finance

Examples · Global

What is alternative data in finance? Examples from car parks to card panels

What is alternative data in finance? Examples make it concrete: satellite images counting cars at Walmart stores to estimate revenue, anonymised card transactions forecasting a retailer's quarter, scraped online prices measuring inflation daily, app downloads tracking a fintech's user growth, job postings signalling expansion weeks before guidance, ship-tracking data for commodity flows, mobile foot traffic for restaurant chains, corporate jet flights hinting at deals, congressional stock disclosures, product reviews for quality problems, employee reviews for culture risk and bank-statement data for lending decisions. Each example below explains the signal, the users and the limits.

What is alternative data in finance? Examples from car parks to card panels

Example 1: counting cars at Walmart from space

The most common search on this topic is simply what is alternative data in finance examples, and the oldest answer is this one.

In 2010 Remote Sensing Metrics LLC used satellite images of a representative sample of 100 Walmart car parks, counted cars month by month, and modelled quarterly revenue from the traffic. It does not give a dollar figure, but with a calibrated model it gives a revenue estimate weeks before Walmart reports. The approach generalised to factory activity, oil storage tank shadows, mine stockpiles, crop health and construction progress. Vendors: Orbital Insight, RS Metrics, Planet Labs' analytics partners. Limits: cloud cover, sample representativeness, and the fact that everyone now has the same images.

Example 2: card-transaction panels forecasting retail earnings

Aggregated, anonymised credit and debit card data from a panel of several million consumers, scaled to the population, tracks spending at named merchants weekly. Analysts use it to forecast same-store sales and revenue for retailers, restaurants, airlines and subscription businesses, and to measure the effect of price changes and promotions. Vendors include Consumer Edge, Earnest Analytics, Facteus, M Science and YipitData. This is the largest category of alternative data spending. Limits: panel bias (cards skew to certain demographics), merchant-name mapping, and the growing share of spending on platforms that card data cannot see inside.

Illustrative: card-panel revenue index versus reported revenue for a retailer (index, 100 = prior year)
  • Q1 panel104
  • Q1 reported103
  • Q2 panel109
  • Q2 reported108
  • Q3 panel101
  • Q3 reported102

Hypothetical figures to show how a well-calibrated panel tracks reported results within a point or two.

Example 3: scraped prices as a daily inflation gauge

The Billion Prices Project at MIT, commercialised as PriceStats, scrapes millions of online prices daily to produce inflation indices that lead official CPI by weeks. The same technique tracks a single retailer's pricing, discounting and stock-outs. Users: macro hedge funds, central banks, retail analysts. Limits: online prices differ from in-store; coverage of services is thin.

Example 4: app downloads and usage for digital companies

Sensor Tower, data.ai and Similarweb estimate downloads, daily active users and time spent for mobile apps and websites. For a fintech, a streaming service or a marketplace, those figures map closely to user growth and revenue. Hedge funds used app data to call turning points in companies from Snap to Robinhood ahead of reports. Limits: estimates not counts, platform policy changes, and web traffic that does not convert to revenue.

Example 5: job postings as a leading indicator

A surge in engineering postings in a new city lands on a job board weeks before it appears in guidance; a hiring freeze shows up as postings disappear. Thinknum, LinkUp and Revelio Labs track postings and headcount by company and by role. Users: equity analysts, macro economists (aggregate postings track labour demand), private-equity diligence. Limits: postings are intentions, not hires; some companies post evergreen roles.

Example 6: ships, ports and customs manifests

AIS ship-tracking data shows tanker movements and port congestion; US customs bills of lading (sold by Panjiva and ImportGenius) show which suppliers ship what to which importers. Commodity traders use the first; supply-chain analysts and short sellers use the second to verify a company's claimed sourcing. Limits: transshipment, data lags and gaps in non-US customs data.

Example 7: foot traffic from mobile location data

Anonymised location pings from mobile apps, aggregated by Placer.ai, Advan and SafeGraph, count visits to stores, restaurants, malls and venues. During and after the pandemic they became the standard way to track reopening by chain and by city. Limits: panel representativeness, privacy regulation that keeps tightening, and visit counts that do not reveal spend.

Example 8: corporate jets, government contracts and congressional trades

Flight-tracking of corporate aircraft has flagged merger talks when one company's jet lands at another's headquarters. Federal contract awards, searchable on public portals, signal revenue for defence and IT contractors. Members of Congress must disclose their trades, and services such as Quiver Quantitative and Capitol Trades publish them; academic work finds mixed evidence of outperformance, but the data are free and widely followed. Limits: small samples and heavy media attention that erodes any edge.

Example 9: product reviews and employee reviews

Review counts and ratings on retail sites track product launches and quality problems; a cluster of one-star reviews mentioning a defect can precede a recall. Employee reviews on Glassdoor and Indeed, aggregated, correlate with retention and culture risk. Natural-language models turn both into scores. Limits: fake reviews, selection bias and platform moderation.

Example 10: news, social media and earnings-call tone

Dataminr detects breaking events from social media minutes before newswires; RavenPack and others score news sentiment by company; NLP on earnings-call transcripts measures management hedging and evasiveness. Users: event-driven and quantitative funds. Limits: noise, manipulation and crowding.

Example 11: bank-statement data for lending

The lender's version of alternative data. Online business lenders and cash-advance funders connect to a borrower's bank account through aggregators such as Plaid or read statements with tools such as Ocrolus, scoring average balance, deposit frequency, overdrafts and existing debits instead of tax returns and credit scores. Consumer lenders add rent, utility and telecom payment histories for thin-file borrowers. This is why alternative lenders decide in hours; the eligibility checker applies the same variables.

Estimated eligibility by product
ProductApproval oddsTypical sizing
Merchant cash advanceLikely$48,000
Short-term loanLikely$40,000
Revenue-based financingLikely$80,000
Invoice factoringUnlikelyNot applicable
Equipment financingLikelyUp to 100% of equipment cost
Business line of creditLikely$60,000
Business term loan (online)Likely$120,000
Bank term loan / SBAUnlikely$168,000

Estimates based on common underwriting thresholds, not a credit decision. Funders also weigh industry, state, deposit consistency and liens. Sizing is a rule-of-thumb ceiling; most offers come in below it.

Full version with instructions: funding eligibility calculator.

Example 12: ESG, emissions and climate exposure

Satellite methane detection, supply-chain audits, weather and flood models and emissions estimates feed ESG scores and physical-risk models used by asset managers and insurers. Limits: inconsistent methodologies and a political backlash in parts of the US market.

How to evaluate an example before you trust it

  1. Does the signal lead the official number, and by how much? Measure the lag in days against reported figures over several periods.
  2. How well does it track? A card panel that lands within one to two points of reported revenue is usable; one that misses by five is noise.
  3. Is the sample representative? Card panels skew by demographic; app panels by platform; satellite counts by weather.
  4. Is it crowded? The Walmart car-park trade stopped working once everyone had it.
  5. Is it lawful and consented? Provenance, privacy and terms of service, reviewed before use.
  6. What is the cost per unit of edge? A $2 million panel that improves a quarterly estimate by one point may or may not pay for itself.

What the examples have in common

Alternative data examples summarised
ExampleTypePredictsLead time
Car-park satellite countsImageryRetail revenueWeeks
Card panelsTransactionsRetail and consumer revenueWeeks
Scraped pricesWebInflation, pricingDays
App downloadsApp usageDigital user growthWeeks
Job postingsOperationsHiring, expansionWeeks to months
Ship tracking and manifestsLogisticsCommodity and supply flowsDays
Foot trafficGeolocationStore visitsDays
Jets, contracts, congressional tradesPublic recordsDeals, revenue, sentimentDays
Product and employee reviewsTextQuality, cultureWeeks
News and call toneTextEvents, sentimentMinutes to days
Bank-statement dataTransactionsBorrower repaymentReal time
ESG and climate feedsMixedRegulatory and physical riskMonths
  • Every example measures behaviour rather than reports, which is why it leads.
  • Every example needs a model to turn the raw count into a financial estimate; the data alone is not a forecast.
  • Every example decays: once the dataset is widely sold, the edge compresses, which keeps vendors hunting for new sources.
  • Every example carries a compliance question: provenance, privacy, MNPI, scraping terms.

Run the numbers

Questions people ask

What is the most common example of alternative data in finance?

Aggregated credit- and debit-card transaction data used to forecast retailers' and consumer companies' revenue; it is the largest category by spending.

What is an example of alternative data in lending?

Bank-transaction data read through an aggregator to score a small business's deposits, balance and overdrafts instead of relying on a credit score.

Is satellite data still useful for investing?

Yes for commodities, logistics and construction, less so for retail car parks, where the signal is widely available and online sales have grown.

Are congressional trading disclosures alternative data?

Yes. They are public records outside company reporting, published by several free services, and widely followed, though evidence of an investable edge is mixed.

Sources and further reading

Related topics

Updated October 12, 2026 · By Alternative Finance editorial team

Next step

Compare real offers from funding partners

Tell us a little about the business and we route the request to partners that fund your industry, size and credit profile. You compare the total payback and decide. Nothing is binding until you sign.

  • Soft inquiry to request offers; a hard pull only if you proceed with a funder.
  • Decisions in as little as 24 hours for many products; funding often within days.
  • Use our calculators to check the true cost before you accept anything.

See what you may qualify for

Takes about a minute. Requesting offers does not affect your credit score. Free, no obligation.

We are not a lender or broker. Requests are forwarded to funding partners free of charge. How the site is paid.