Unlocking Hidden Data: Kernel Provides the Key
The future we imagine…
Consider for a moment a Bloomberg Terminal — a nexus for hundreds and thousands of signals that collectively paint an accurate, near real-time portrait of global financial markets. These signals are delivered reliably to hundreds of thousands of professionals, helping them create opportunity, hedge risk, and find efficiency. So, why isn’t there a Bloomberg terminal for Earth?
What would we do differently if we could trace even small changes in plant health across millions of acres of corn as accurately as we now trace shifts in the global commodities market? What if we could quantify and monitor the stock of carbon sequestered in our forests with the same confidence with which we now follow the New York Stock Exchange? Or monitor the world’s food supply on a daily basis?
This work isn’t easy. We assemble whole-Earth image stacks that produce information twice daily over 14 years and across 32 spectral bands — that’s more than 4 quadrillion data points (and counting). Because so few people wrestle with this data, accessing it is pretty tough — the archives are difficult to navigate and manipulate, tools and methods are missing or sub-scale, and data products have quirks and gaps that need filling. Virtually all of the effort is driven by science teams who are sharply isolated from each other and from industry. Thus, this vital information can often feel locked away: inaccessible and unclear.
Why does it matter?
We are convinced that accurate, timely, and localized coverage of our planet will change what’s possible in tackling some of our toughest collective social and economic challenges surrounding food scarcity and security. At TellusLabs, we are committed to unlocking this data from overhead as an asset to everyone: from large agribusinesses to individual traders, we want to make this information accessible for you. Through Kernel, the data is yours: how will you use it?
At TellusLabs, we have assembled a small team of data scientists, geographers, and engineers who work closely to unlock that stranded data as a unified team on a unified technology stack. Together, we’re assembling tools that collectively represent a simple, efficient, repeatable system for inventorying, monitoring, and forecasting the health of our fields, forests, and water. Doing anything reliably on a daily basis and across the whole surface of the Earth is audacious. Thanks to our advances in remote sensing, cloud computing, and machine learning, our models are becoming more and more accurate:For the past two years, their agricultural insights product, Kernel, delivered best-in-class results for the largest harvests in the world. By producing information more accurately, frequently, and quickly than our competitors, we are making spaceborne insights visible on Earth. While this data may still be locked away, our software Kernel provides the key.
