A business centric approach to fast queries on Data Lakes.

Slice and dice analysis on ,multi-dimensional data

Use SQL and Python to power your B.I and analytics with 100 times the speed of comparable in-memory technologies.

Advanced optimizations

SNAP includes advanced optimizations that allow you to use complex SQL joins and still achieve sub-seconds response times for large datasets. SNAP QuBEs define a logical data model of your business domain which can then be reused across multiple applications and front ends. QuBEs can be modeled to take advantage of advanced optimizations for queries involving timestamps, metric binning, high cardinality dimensions and more.

Machine learning and A.I analysis on SNAP

Connect Python, R, Scala to SNAP and optimize your data science workloads. Use all your data instead of a subset of your data.
Combine SQL with Python/Scala or R at blazing speed.

No Extracts

Eliminate building summary tables, extracts and specialized aggregations for each use case. Store your data in Sparkline at the lowest granularity and aggregate on the fly.

Compressed data and indexes

Sparkline uses compressed in-memory data and indexes to accelerate your queries. For many business intelligence type queries, just in-memory technologies are not adequate.

What works for reporting, does not work for analytics

B.I is different than just reporting and dashboards

There are two kinds of use cases for data. One is operational reporting and dashboarding. The objective is to give a snapshot of business metrics as of a time and allow users to see changes over a short window.
Fast, True B.I which is where analysis plays a big part is an exercise to dig into the data and search for patterns and trends. Here the emphasis is not snapshots but finding trends across various dimensions.

True B.I requires a more sophisticated platform.