About Atlas Analytics
Atlas Analytics conducts macroeconomic forecasting from outer space. In short, we make predictions about GDP and the stock market using satellite imagery. Follow us for economic and financial commentary, news and forecasts.
Atlas Analytics Overview Video
Satellites circumnavigate Earth every minute of every day. Harnessing the power of these elements constantly taking pictures of Earth’s surface, our team has developed an algorithm that extracts information about the built environment and transforms this into numerical data.
Then comparing this satellite data to past GDP, Atlas Analytics has created an AI and machine learning model to use previous patterns to predict future economic activity. As the satellites continue overhead, we’re able to use remote sensing to model real-time GDP to improve knowledge about the health of the world's economies in real-time.
Currently, Atlas Analytics’ model forecasts GDP for the United States. This is extremely valuable information as the US Bureau of Economic Analysis (the official government agency tasked with reporting US GDP) only reports this statistic quarterly with up to a three month lag. For our next iterations, we plan to expand this technology to other major economies and to regions that don’t explicitly report their GDP data.
How does Atlas Analytics Work?
Source: iStock
Atlas Analytics’ algorithm ROY (Remote Orbital Yield) uses a four-step process to predict GDP.
Ingest Satellite Imagery: First, ROY pulls the most recent Landsat satellite images and appends them to the images already in the algorithm’s database. The Landsat Program is a joint venture between NASA and the US Geological Survey (USGS) that has been orbiting satellites around Earth since 1972. Using remote sensing, we then apply a custom algorithm to extract numerical data about the built environment. As the change in the built environment (e.g., construction) is a strong proxy for economic growth, this is our key variable for forecasting GDP.
Pull Traditional Economic Data: Using APIs connected to government data, Atlas Analytics pulls in real-time traditional macroeconomic statistics.
Collate and Clean The Data Pipeline: Combining the numerical satellite imagery with the traditional economic statistics, we move the data through our data pipeline, utilizing feature engineering, imputation and seasonal adjustment to prepare a final, prediction-ready dataset.
Employ AI and Machine Learning to Predict GDP in Real-Time: Finally, we build a statistical learning model relating nearly 40 years of feature data (including, in particular, the satellite imagery) to past GDP growth rates. Utilizing machine learning and AI, we then predict current GDP ahead of government releases.
How was Atlas Analytics Developed?
Atlas Analytics’ algorithm is the brain child of Jake Schneider in 2018 and was developed by Michael Trotta, Paige Kirby and Ryan Krueger between 2020 and 2022. By the spring of 2022, our team had completed the primary development of the algorithm and since then we’ve been monitoring the results. Now convinced of the efficacy of our invention, we are sharing the results with you, the investing public.
About the Inventor
Jake is the creator of Atlas Analytics and a current Senior Economist with the City of Boston’s Planning Department. Previously, he’s served in similar roles at Bain & Co, The World Bank, The Brookings Institution and Alan Greenspan’s consulting firm. He holds graduate degrees from Harvard University and Oxford University, and an undergraduate degree from The College of William & Mary.
Want to Trade Atlas Analytics?
Once a month, Atlas Analytics posts our proprietary GDP forecasts and price targets for four macro-exposed ETFs to make you a better-informed trader. Want access, visit our Client Portal page or subscribe below.
