Non-traditional datasets - satellite imagery, web activity, card transactions, geolocation - used to form a view before it reaches the mainstream.
Alternative data ("alt data") is any dataset that sits outside the conventional inputs of company filings, exchange price data, and sell-side estimates. The canonical examples are satellite imagery of parking lots and ports, web-scraped pricing and hiring, anonymized card-transaction panels, and mobile geolocation. Funds pursue it because it measures activity directly, ahead of the reported number that only describes it later.
The value of an alt-data set decays as more people use it, so freshness and differentiation matter as much as accuracy. Raw alternative data is also noisy, uneven in coverage, and full of panel biases that have to be corrected before it is usable. And it carries a compliance dimension that traditional data does not: a dataset has to be properly licensed and free of material non-public information, which is why sourcing and documentation are part of the product, not an afterthought.
A live evaluation measures EdgeOrigin against your coverage requirements: decision-ready real-time data, delivery latency into your systems, and record-level provenance.