The application of satellite data and alternative datasets for fundamental analysis

Let’s be honest. For decades, fundamental analysis was a game of patience and paperwork. You’d pore over quarterly reports, listen to earnings calls, and try to read between the lines of management commentary. It was like trying to judge the health of a forest by only looking at a single, carefully pruned tree in the visitor’s center.

Well, the game has changed. A quiet revolution is happening, and it’s happening from space—and from a thousand other digital breadcrumbs we leave behind. Today, the sharpest analysts aren’t just looking at balance sheets. They’re counting cars in parking lots, measuring shadows cast by oil tanks, and tracking global ship movements in real-time. This is the new frontier: the application of satellite data and alternative datasets for fundamental analysis.

What exactly are we talking about here?

First, a quick sense of the landscape. “Alternative data” is a broad term. It basically means any information generated outside a company’s official channels that can offer an investment edge. Think of it as the digital exhaust of the global economy.

And satellite imagery is perhaps the most literal “big picture” dataset of them all. But it’s far from the only one. Here’s a quick, non-exhaustive list of what funds are using now:

• Geolocation data: Foot traffic metrics from mobile phones for retailers, restaurants, and malls.
• Satellite & aerial imagery: Monitoring agriculture, mining activity, construction progress, and even the darkness of roofing tar (a weirdly good indicator of new home construction).
• Web scraping & social sentiment: Tracking product reviews, pricing changes, and brand mentions across the web.
• Credit card transaction data: Aggregated, anonymized purchase data to gauge consumer spending trends.
• Marine & aviation tracking: Following global supply chains via ship AIS signals and plane flights.

From space to spreadsheet: how it actually works

Okay, so you have this firehose of data. The real magic—and the hard part—is turning pixels and pings into a tradable insight. It’s not about pretty pictures; it’s about creating consistent, quantifiable metrics.

Case in point: the retail reckoning

Imagine you’re analyzing a big-box retailer. Sure, you can wait for their monthly sales figures. Or, you could analyze satellite imagery of their store parking lots over time, comparing year-over-year occupancy. Even better, you could blend that with geolocation data from smartphones to see not just how many people visited, but how long they stayed.

A dip in average visit duration before an earnings report? That might signal customers are having trouble finding what they need, or are just browsing—a potential leading indicator of weak same-store sales. It’s a tangible, unfudgeable data point that exists outside the company’s narrative.

The supply chain crystal ball

Here’s another powerful application. After a major hurricane, how do you accurately assess the impact on oil refineries or chemical plants? Official updates can be slow. Satellite data, however, can detect heat signatures from flares (indicating operational status) or measure water levels around facilities. Investors used this exact method during recent Gulf Coast hurricanes to model disruptions and predict commodity price swings.

Similarly, tracking the volume of ships sitting outside major ports like Shanghai or Los Angeles gives a real-time, unvarnished view of global trade logjams—information invaluable for analyzing shipping companies, manufacturers, and retailers alike.

The challenges? They’re real.

This all sounds like a silver bullet, right? Not so fast. Applying alternative data comes with its own set of headaches.

First, there’s the signal-to-noise ratio. You’re drowning in data. Is a 5% drop in mall traffic a trend, or just bad weather last Tuesday? Isolating a clean, predictive signal requires sophisticated data science and a lot of back-testing.

Then there’s interpretation. More cars at a car dealership lot could mean booming sales… or a glut of unsold inventory. You still need the fundamental context—the human analyst’s brain—to make sense of the numbers.

And let’s not forget cost and access. Some of these datasets are incredibly expensive, creating a potential “data divide” between large institutional players and smaller firms. The edge, frankly, is getting crowded.

The future is integrated (and a little bit messy)

So where is this all heading? The real power won’t come from any single dataset. The killer app is data fusion—layering satellite imagery with geolocation pings, social sentiment, and traditional financials to build a multi-dimensional, living model of a company.

Think about analyzing a farm equipment manufacturer. You could combine:
1. Satellite data on crop health to predict farm income.
2. Geolocation data at dealerships.
3. Online forums where farmers discuss equipment purchases.
4. The company’s own inventory and credit data.

Suddenly, you’re not just reacting to an earnings miss. You’re anticipating it quarters in advance. You’re seeing the story unfold on the ground, in the fields, and in the conversations long before it hits the official financial statement.

That said… this isn’t about replacing the fundamental analyst. It’s about augmenting them. Giving them super-senses. The core questions remain: Is this a good business? Is it run well? Is it cheap? Alternative data just provides harder, faster, and sometimes deeper evidence to answer those age-old questions.

The forest is no longer a mystery. We now have satellites showing us the canopy cover, sensors reading the soil moisture, and audio devices tracking the animal movements. The trick, now, is learning to listen to all of it at once—and to understand what the symphony actually means.