Skytree App:
Skytree Server is an enterprise-grade software platform for developing, testing, and deploying advanced analytics solutions for big data.

Get the App

Learn More

Skytree and MapR: Bring Machine Learning to Life in the Enterprise


Bringing Machine Learning to Life in the Enterprise

Data Science Principles, Game-Changing Predictions

Machine learning involves the creation of algorithms that allow for the analysis of massive data sets that traditional BI tools cannot handle. The more data that is analyzed over time, the smarter and more accurate the algorithms become. It’s an iterative learning process of continuous improvement. So whether you are an online dating company seeking to best match couples and personalize pricing scenarios, a media company matching buyers and sellers through ad arbitrage, or simply measuring distances to galaxies, running Skytree on top of your MapR Apache Hadoop platform will rapidly deliver big answers to your big data questions.

Enterprise-class predictive modeling across massive data sets requires a scalable, integrated architecture that combines complex, self-improving algorithmic capability with a reliable, robust and dependable Apache Hadoop platform— highly available, 24x7.

The Skytree and MapR solution offers exactly that—a potent and compelling big data platform to reliably underpin the iterative processes native to machine learning.

Why Skytree on MapR is Unique

MapR provides several key advantages to make analytics professionals more productive. MapR snapshots and volumes allow users to build and test models on the same cluster for production data without impacting operations. This also allows for easy versioning of models and back testing against historical data sets. In addition, unlike other distributions for Hadoop, only MapR provide a fully read-write data platform which allows existing applications, custom libraries, and modeling languages, and scripts (e.g., Grep, Git) to work out of the box. Moreover, data movement is quick and easy with MapR Direct Access NFSTM without requiring a separate cluster for data ingest.