POWER > Systems Management > Data Management

Three Analytics Trends Affecting IBM i

IBM i Analytics


The technology world isn’t slowing down. Cloud is mentioned everywhere, and advances in analytics continue to emerge almost daily. We’re hearing about the Internet of Things (IoT), reading about big data and talking about IBM Watson* technology. The data scientist role is becoming more prominent in enterprises to focus on data analysis. And yet, many IBM i shops consider themselves stuck with antiquated reporting tools and incomprehensible data environments.

For a company running a majority of its business on IBM i, and storing a majority of its operational data in Db2* for i, you might be asking if you should be taking advantage of this evolution. In this article, we’ll provide some thoughts on three trends around the evolving analytics world—empowering data analysts; accessing services in the cloud to augment operational data; and automating and repeating data preparation, consolidation and feeds into and out of the cloud—that may be of value for IBM i clients.

Empowering the Analyst

One trend we’ve noticed in the industry is the transition of who’s using business intelligence. In the early days of Query/400, the report author was someone in IT with a deep knowledge of the data (and for many of you, this is still true today). End users are simply receiving the data—and in some cases, need to do additional number crunching to analyze it.

Several years ago, a metadata layer approach, which offers a way to shield report authors from the complexities of the underlying database, emerged. By implementing a solution with a metadata layer, you’re shifting the role of building reports to advanced end users outside of IT and creating an enterprise approach to data access (ensuring security, standardization and “single version of the truth” concepts). The business analyst job title has emerged, and the metadata approach enables them to spend more time analyzing data without a dependency on IT.

An even newer job title has been appearing lately: the data scientist. Part of this is because of the rise of unstructured data (e.g., tweets). While this is mostly happening in larger enterprises, even small- and medium-size businesses are looking for data-centric analysis or “data discovery.”

What is data discovery? Think of it this way: Traditional report writing consists of laying out what data elements will be seen by the consumer of the report. With data discovery, all data elements are available to the data analyst in real time so they can dynamically add, remove, filter and immediately visualize trends to glean “ah-ha” moments from the analysis.

A pitfall of this new trend is that it can leave the user in a state where they are crunching the data more often than analyzing it. Data discovery can also leave individual analysts with different “versions of the truth.” An effective data discovery effort needs to be integrated into the reporting metadata layer. An example of this kind of integrated approach is the new InfoAssist+ feature of Db2 Web Query that allows you to leverage its metadata and access capabilities instead of offering a completely independent tool for that function.

Accessing Cloud Services

IBM, in conjunction with COMMON, recently hosted an inaugural workshop called “IBM i Driveway to Watson–Rev Up Your Business Applications.” The goal of the session was to provide real examples and hands-on labs around integrating IBM i applications and data with Bluemix*, which is IBM’s set of interfaces in the cloud, including Watson. If you’re unsure where to start, this session was a good way to understand some of the possibilities.

IBM’s Db2 Web Query team participated in this event by showing a prototype around weather-related sales analysis.

Through IBM’s Bluemix interface, you can access Weather Company data from an IBM i application through a set of web services and weather service APIs. The prototype application shipped store location data in our operational database up to the cloud-based service on a daily basis, and the weather service returned information about forecasts for that day and location in JSON format. Using SQL and Db2 functions such as the new JSON Table function, data returned in JSON format was shredded into relational rows and columns and then combined with operational data of interest in a data mart. The data mart then provides the repository for analyzing data across various weather attributes.

Data Preparation and Automation

A critical reason to consider using Bluemix cloud data is the ability to make data gathering automated and repeatable. Some plumbing is required to facilitate the consolidation of data and loading of the sales data mart.

The automated daily process looks like this:

  • Cleanses/prepares data to feed up to the Bluemix service
  • Combines operational data with the daily weather info
  • Updates a dashboard with weather-related store sales analysis

With the process well-defined and some simple SQL programs, views and metadata in place, the DataMigrator component of Db2 Web Query provides the plumbing to automate the process on a daily basis. While the data mart now enables weather-related analysis, it also represents an untangled set of operational data that can be the source to push into other Bluemix/Watson services for additional insights.

In the end, the ability to leverage advanced visualizations, access services in the cloud and manage the data plumbing for this are all capabilities that exist today, thereby reducing risk, accelerating time to value and leveraging what you do best: IBM i.

Doug Mack is an analytics consultant in the DB2 for i Center of Excellence at IBM.

Gary Goldberg is the Vice President for WebFOCUS System i Products at New York-based Information Builders Inc. and is responsible for the delivery of DB2 Web Query for IBM i product sold by IBM.


comments powered by Disqus

Advertisement

Advertisement

2017 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

POWER > SYSTEMS MANAGEMENT > DATA MANAGEMENT

IBM Researchers Maximize Apache Spark’s Capabilities

The Evolution of Domino Attachment Object Service

DAOS changes mean easier administration, backup and resync

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store