Dataiku the maker of Data Science Studio (DSS), announced today the integration with the advanced data processing engine, Apache Spark. By adopting Spark, data analysts can process much larger Hadoop data sets, ranging into the terabytes and also process that information much more quickly.
Paring the capabilities of Apache Spark with the advanced analytics features of DSS creates significant opportunities for those looking to leverage very large data sets. DSS provides an IDE (Integrated Development Environment), which gives developers the tools to rapidly build Ad Hoc queries, which are then processed against selected data sets, creating visual representations of the relationships found in the data.
Visual Recipes, which are a core component of DSS, can now be executed on the Apache Spark framework, while leveraging the SparkSQL programing language and data processing engine. That helps DSS users perform tasks such as joins and aggregations dozens if not hundreds of times faster than what could be accomplished with Hadoop using Apache Hive.
Apache Spark integration also gives DSS the ability to work with Spark R, SparkSQL, and PySpark, which brings R, SQL, and python based programing to the Spark environment. Much like the other components of Spark, PySpark and Spark R eases and speeds the native capabilities found in DSS and makes Spark a viable alternative to the traditional Hadoop/Hive stack, while also allowing analysts to share data engineering recipes and limit the need to recode or redevelop algorithms.
The integration of Apache Spark brings with it many other advantages, all of which dovetail well into the inherent capabilities of DSS. Those advantages include:
- Data Volume: Spark enables data analysts to use DSS to deploy advanced algorithms across several hundred gigabytes of data.
- Collaboration: The PySpark and Spark R frameworks makes it easier for team members to share cluster resources.
- Education: DSS offers a unified interface for multiple frameworks, allowing users to immediately delve into the capabilities of Apache Spark, without having to learn the intricacies of a new of technological frameworks and dialects / languages.
- Future Proof: Thousands of contributors are continually working to enhance the Spark Project, creating new standards and enhancements, which are rolled into the DSS/Spark environment.
Another important element that DSS brings to the table with Apache Spark is the ability to train models using both MLlib and Scikit-Learn. By adding MLlib to the mix, users are now able to address large scale projects by being able to model the data sources in their entirety. That in turn allows analysts to leverage the full cluster of data services and avoid the problems normally associated with a “divide and conquer” approach that may miss important segments of data.
The addition of Apache Spark to the extensive number of datastores already supported by DSS, allows analysts to create large scale big data analytics projects, without the risk of reaching beyond the capabilities offered by data engines currently in use.