Building Data Architectures for BI and Analytics

Bookmark and Share

Before enterprises create data architecture for business intelligence and analytics, there are a few key components they need to consider.

Richard Sherman, founder of Athena IT Solutions, took part in a Data Summit 2015 session titled, “The Future of Data Warehousing” where he spoke about creating the right conditions for business intelligence and analytics.

Data Summit 2015 took place May 11-13 at the New York Hilton Midtown.

Many enterprises are struggling with a “data deluge” Sherman pointed out. This constant inundation of data comes from society in the forms of social media, news, and constant connectivity through smart phones and tablets.

“[Nowadays] IT and businesses don’t control some of the data that they are using,” Sherman said.

There is no “silver bullet” solution but there are some steps businesses need to consider before jumping to a platform to help handle their large swaths of data.

Businesses must consider the industry and size of their enterprise along with their employees’ roles as to who oversees the data or what purpose the data is being used for, Sherman said. Is it for operational or managerial purposes?

Enterprises must take into account the different analytical skills they are working with, whether it’s for consumers or producers.

Consumers have a wide range of analytical needs, according to Sherman.

After factoring in those considerations, enterprises need to look at the different data sources and types they are dealing with.

“The key is to get master data management and conform data that is consistent in the data warehouse but not apply business rules or algorithms or metrics until you get to the data marts,” Sherman said.

Enterprises can then begin building successful data architecture by tapping into a variety of systems technologies including systems of record and data capture (SOD), systems of integration (SOI), systems of analytics (SOA), and systems of business intelligence and analytics.

Examples of SOD include enterprise applications, CRM, web services, cloud applications and databases, business processes, data services, social media, unstructured data, big data databases, and the Internet of Things.

SOI includes staging, integration and distribution schema, operation data store, and master data management while SOA uses data marts, OLAP cubes, columnar, data science hubs, analytical sandboxes, or refined big data.

After setting up the architecture, businesses can tap into in memory columnar, OLAP cubes, Power Point Excel services, or BI repositories to do reports, dashboards, data visualization and discovery, spreadsheets, big data analytics, mobile BI or predictive analytics, Sherman explained.

To access Data Summit 2015 presentations, go here.