Big Data-Driven Alternative Economic Indicators
Analyzing high-frequency data to provide real-time insights into the economy.
Retail Activity vs. Official Retail Sales
Methodology: Satellite Imagery Analysis
The Retail Activity Index is generated by using machine learning algorithms to analyze satellite images of over 5,000 major retail parking lots across the United States. By counting the number of cars on a weekly basis, we create a high-frequency proxy for in-person retail foot traffic. This leading indicator often detects shifts in consumer behavior weeks before official monthly government data is released.
Online Hiring vs. Official Unemployment Rate
Methodology: Web Scraping & NLP
The Online Hiring Index is built by scraping millions of job postings from dozens of corporate career sites and job boards daily. Natural Language Processing (NLP) is used to de-duplicate and classify new job openings. A significant increase or decrease in the volume of new postings provides a real-time signal of labor market strength or weakness, often preceding changes in the official unemployment rate.
Consumer Spending vs. Official GDP Data
Methodology: Aggregated Transaction Data
The Consumer Spending Pulse is derived from anonymized and aggregated credit and debit card transaction data. By tracking spending across various sectors on a weekly basis, this index provides a timely view of consumer health. It serves as a powerful nowcasting tool for Personal Consumption Expenditures (PCE), the largest component of quarterly GDP.