alphaspirit - Fotolia
Maine's analytics capabilities were stuck in the last century until just months ago, but by adopting Oracle Analytics the state government has transformed its business intelligence capabilities.
The database the Maine state government used up through part of 2019 was built in 1995, said Jeffrey Jordan, director of enterprise data services for the State of Maine, during the Oracle Analytics Summit, the vendor's virtual user conference.
The state's data was deeply siloed, and while users in a given department might know where to find that department's data, if they wanted data from another department they didn't know how to access it.
"That created a level of complexity for the end users in state government that was unmanageable," Jordan said during a session recorded on Tuesday. "You were left with an environment where the average data consumer in state government only knew how to get what their predecessor knew how to get. They may have been trained to run a certain report, but they didn't know how to get any other information."
Meanwhile, anyone who needed to build a report or dashboard had to go through the state's IT department to get it, and that took time. Users needed to submit a ticket, which then went into a queue, and the IT department would get to it when its turn came up.
Maine's path to Oracle Analytics began by making a case for change to the leaders of the state government. Jordan and his team of database analysts and IT professionals wanted to make the state government's BI capabilities not only faster but also accessible by more than only the members of the IT department.
That meant first migrating the state's data to the cloud, and then completely revamping the analytics process.
"A quote by one of the people I interviewed [about the need to modernize Maine's BI technology] has stuck with me forever," Jordan said. "They said, 'When a customer asks me a question, they've lost interest by the time I can provide an answer.' That was a matter of function, and the availability of the data that was needed."
Once given approval, Maine's analytics modernization began in 2018 by remodeling its data sets and moving them into an Oracle Analytics Cloud data warehouse, a process that lasted through the end of 2019. Meanwhile, Jordan and his team determined they wanted to build a three-tiered service model using Oracle Analytics.
First, the team designed a data analytics workbench to enable data analysts to easily develop the reports needed to operate the state government and do so with the modern visualization tools they were previously lacking. Next, they built an information portal including natural language processing capabilities to make dashboards and queries immediately available to members of the state government outside Jordan's team, the business users.
Jeffrey JordanDirector of enterprise data services, State of Maine
Finally, a data science lab where machine learning can be applied to data and forecasting models can be developed, is still under construction.
In addition, leaders in state government wanted a single point of entry to the data, eliminating siloed data and that inability to access data in other departments. In response, Jordan's team built an information portal using Oracle Analytics where end users can find what they're looking for using simple drop-down menus.
"We wanted to make it quick and easy to get to the data you needed to do your job," Jordan said.
Beyond simplifying -- and speeding up -- the analytics process to make it accessible to more users, Jordan added that implementation of Oracle Analytics and the transformation of the Maine state government's BI stack has improved partnership between IT and end users.
"Answers are delivered faster, and there's a reduced reliance on IT," Jordan said. "It doesn't go to zero, but it's reducing the reliance so that you, as a data consumer and data producer, work with your data without having to wait in the queue."
By May 2020, the state's reliance on the legacy system was nearly eliminated and the old system is now used only to make sure reporting is consistent through the end of the year.
"What we're seeing is a high value to end users," Jordan said. "They're turning what used to be one-off processes that they had to do manually into something the system does for them. They set it and forget it, and just get the information into their hands on a regular basis, and they're seeing a lot of efficiency gain."
Despite the success of Maine's data transformation, the transition from its legacy stack to Oracle Analytics hasn't been completely without glitches. Those problems, however, have more to do with education and training than anything else, Jordan said. As a result, more training is being planned.
In addition, Jordan said that he regrets not using Oracle Analytics until after the data was remodeled. The platform, he said, could have been used to do ad hoc analysis had training been done earlier, and by doing one step and then the other it delayed the starting point when the state government began getting value out of Oracle Analytics.
Meanwhile, as Jordan and his team continue developing the state of Maine's analytical capabilities, they plan to move the state's data from Oracle's database as a service to Oracle Autonomous Data Warehouse and integrate the state's human resources data into the warehouse.
"We're really excited to have replaced the old," he said. "Now, we maximize the value we're getting out of this new platform."