World’s largest AI Data Science Feature Superstore

“Pre-pandemic data sets and assumptions won’t fly in the post COVID-19 world.
Businesses are starting over, and they’re going to need to understand …”

For Data Scientist Buyers

Our data features superstore allows you to quickly discover, evaluate and purchase from our inventory of over 100,000 and growing external features

Learn More

For Data Sellers

Tap into the exploding AI data science market

AI data science represents a multibillion-dollar market
for data growing at 50%/year

Learn More is a cloud-based data science features platform that makes it easy to find and use actionable external features

Platform and Marketplace Overview

Mashup 1000s of Data
Signals in Any Format

We have a universal semantic layer that describes
every DataTile, where it came from, how it was
calculated, what its quality is, and how it connects
with every DataTile. This provides unprecedented
ability to search for relevance, perform statistical
analysis across hundreds of thousands of
data signals, and mashup thousands of data signals into
any format or grain data set for analysis

To find the actionable signals within
this data cloud, Nitrogen offers multiple approaches to
discovering actionable data signals

Use our DataTiles FacetedSearch to Find Data Signals you Need provides both an intuitive and an analytic approach to discovering actionable features

For the intuitive approach, you can use faceted and keyword

search to find the features you know you need

For the analytic approach, you can use correlation and machine
learning to automatically find features that are
most likely to be predictive

You can run hundreds of thousands of correlations against your dependent variable to identity
actionable external features and then use Lasso multivariate regression to reduce
 collinearity and find the best variables for your model.

This joined set of data is delivered in your desired format: CSV files, JSON,
or Parquet to allow Integrate into your current infrastructure and data pipeline.

Human in the loop
Curation into Features

Faceted search to narrow

Correlation of Features to
increase relevance

Use Lasso multivariate regression
to decrease collinearity

Join relevant Features into
analysis ready data frame

This joined set of Features is
delivered in your desired format that can
integrate into your data pipeline

Our Machine Learning
Feature Selection
Funnel shortens the time it takes to find relevant variables. organizes our feature cloud into a universal semantic layer based on time and geography. It defines how features are aggregated, enabling the automatic combining of features that are on different time and geography scales. Sophisticated statistical analysis compares your internal data to external features, and external features to each other, to identify those best for you.

Charts and graphs help you to understand the features and generate insights. Once you’ve selected the external features you want, we continuously deliver them to your data pipeline.

Find your
Actionable External Features

Please contact us for a discussion and demo of the DataTiles platform.

Get Started Here