TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Upside Briefing: Dell Statistica

Dell recently released Statistica 13.1. We offer Upside's first impressions from a business briefing and a look at the idea of a citizen data scientist.

• By Fern Halper
• May 16, 2016

Name of Company: Dell Statistica

Company Location: Round Rock, Texas; Global

TDWI Product Market Categories: Advanced Analytics, Predictive Analytics

Company Vision: By embedding analytics everywhere and empowering a wider community, you'll transform your business and achieve new levels of success.

Briefing Notes: I spoke with Shawn Rogers, chief research officer, Dell Statistica about the release of Statistica 13.1. There are several interesting parts to this release, including support for the "citizen data scientist," new edge analytics for IoT, and Dell's support for analytics marketplaces. I'm writing about the first two here.

The making of the citizen data scientist: Rogers and I spent a lot of time talking about the "citizen data scientist," aka the BI power user. Dell Statistica, of course, provides advanced analytics for data scientists and statisticians. Its goal is to make its products easier to use for the power user. To that end, it announced a number of new features in 13.1 including:

• A new data preparation feature. Dell now provides 80 prebuilt functions for data preparation accessible through a drag-and-drop and point-and-click interface. This enables citizen data scientists (as well as data scientists and statisticians) to access data from multiple sources, such as spreadsheets, third-party data, databases, Hadoop, and even screen scrapes from websites for analysis. Data preparation capabilities include what Dell is calling a "health-check node" to help profile the data to look for issues such as misspellings and duplicates.


• The ability to reuse models via executing an external workspace. This feature allows someone to build a model and then lock it down in an external workspace. The model then becomes reusable. A data scientist might build out the steps for a churn model that a marketer "citizen data scientist" can then execute as part of a work process. This provides a repeatable, standardized workflow.


• Inline visualization as part of the analysis process. Dell has added visualization capabilities to the advanced analytics workspace so users can explore their data as they are iterating on their model building.

Edge analytics:We also discussed Dell's ability to perform edge scoring for IoT. According to Rogers, IoT is becoming increasingly important to Dell. Edge analytics refers to performing analytics close to the edge of a network. Dell has the capability to build models in their gateway as well as deploy models on edge devices to score data at the edge. In fact, Dell Boomi atoms (its integration software) enables transport of these models to different devices in an IoT network, where they can be embedded.

First impressions: There were a couple of things I found interesting in this briefing. First, I liked Dell's positioning around the making of a citizen data scientist. Statistica has been around for a long time and has worked closely with data scientists and statisticians. The tools weren't that "pretty," but end users loved them anyway. Now the tools have a much nicer UI. However, the move to support citizen data scientists is more than an easy UI, and Dell knows that. Dell understands that there is a gap between a data scientist and a business analyst/super user that cannot just magically disappear because of software.

The company is taking a pragmatic approach to helping the citizen data scientist by (1) thinking of this chasm in terms of a strategy that includes software as well as skills and (2) providing a way for models built by data scientists to be reused by citizen data scientists by putting them into an external workspace and a repeatable workflow that provides a way to standardize the models.

Second, I was interested in the fact that Dell enables organizations to embed Statistica at various points in an IoT network including at the edge of that network in devices. This supports different kinds of IoT use cases. For example, if you need to analyze data and react in real time, then it can make sense to put analytics in an edge device. Statistica can do that. I also found the Dell Boomi atoms interesting as a way to transport and embed analytics at the edge.