Survival of the Fittest in the Modern Digital Age

Powered by technology, we live in an increasingly elaborate and fast-moving global society. With trends bubbling and fading as fast as tweets, consumer behavior is ever more dynamic and supply chains are more global than ever before.

In this frenetic data-driven new world order, developing capabilities to disentangle complex relationships and predict key emerging trends is foundational for future success. Governments, businesses, and investors alike must rapidly adapt; those that don’t will simply fail.

Big Data Hunting

There is an abundance of information originating across the globe. Big and small data alike tend to be wild and unwieldy — large in size, and requiring capturing, cleaning, aggregating, retrieving, mapping, and connecting before they can be put to good use.

Humans Alone are Helpless

With big data arise bigger analytical challenges. In this day and age, spreadsheets and simplistic traditional statistical models alone are inadequate for interpreting mass data. Making sense of intelligence to accurately predict business and economic metrics requires advanced and dynamic methodologies.

Leveraging the Machines

For each fundamentals prediction, an algorithm needs to distinguish noise from signals to avoid wrongly specified relationships and spurious correlations. Our algorithms systematically select features and employ machine-learning techniques to train models on historical data. However, with constantly changing global product innovations and consumer tastes, what’s driving an economy or a company’s revenues today might not be the same a year from now. While the best model within an algorithm is selected through rigorous backtesting, faced with new data an algorithm needs to be flexible to switch between models and features. Moreover, they require continuous human improvement in the form of updating, maintenance, and oversight to ensure optimal performance. This ensures a track record of prediction accuracy is built over days, months, and years.

Navigating the Labyrinth

Creating dynamic models to generate predictions is labor intensive and time consuming. Combined with complexities associated with identifying and processing mass data, this explains why some of the world’s largest institutions still find it hard to produce actionable insights and predictions.

We Connect the Dots

Big Data Federation believes the human brain faces limitations in terms of bias and analytical capability and thus the capacity for optimal and objective judgements. Our purpose is to leverage machine- and data-driven fundamentals predictions for profitable investing and decision making.

To this end, our home grown, scalable, and smart data platform continuously amasses and processes tens of billions of data records per day from a diverse and growing number of exogenous data sources.

Our algorithms uncover patterns in these data and identify leading and lagged interconnects to develop fundamentals predictions several quarters out for a company, its customers, suppliers, industry, and the macroeconomy.

Powered By Machine-Enhanced Fundamental Predictions


Billion Data Points Processed Daily


Companies Predicted


Industries Covered


Economies Analysed

Built to Succeed

Each quarter we generate and record predictions for around 1,600 U.S. publicly-traded companies. For the past 16 quarters, our algorithms have consistently predicted business fundamentals with ~99% accuracy.1

1 Prediction accuracy is calculated by comparing actual reported revenues and our predictions 60 days prior to company reporting dates, i.e. it is mean absolute percentage error averaged over predictions made during past 16 quarters.

Accuracy of Predictions