Business Intelligence

Business Analytics and Reporting Pilot (Part 1)

In this web log, I’ll share my experience with the State of Texas Department of Information Resources (DIR) Business Analytics and Reporting (BAR) pilot. During this pilot, Catapult Systems worked with the DIR and five other Texas state agencies to explore their data to gain insights from it. In so doing we determined the suitability of currently available tools for the purpose of a shared production BAR platform in Texas.

Inception

During the 84th Texas Legislative Session, House Bill 1890 said “the Texas Department of Information Resources shall:

Implement a shared data reporting and business analytics service  with appropriate security isolation, for state agencies only. Launch as a pilot program with a limited number of state agencies in order to validate a solution before implementing a statewide service. “

So the DIR released an RFP in December 2015. Catapult Systems competed in the RFP process (which ate my February) and won the bid! The RFP called out fourteen specific capabilities with anticipated need in a business analytics and reporting shared service for Texas state agencies:

  • Dashboarding
  • Search
  • Administration
  • Alerting
  • Self Service
  • Text Analytics
  • Geospatial Analysis
  • Excel Integration
  • Forecast and Statistical Analysis
  • Data Warehouse Automation
  • Big Data Connections
  • Portals
  • Enterprise Service Bus
  • Mobile Capability

Obviously a lot of thought had gone into what capabilities might be needed by these agencies for the foreseeable future. I won’t take the time here to describe each of these capability requirements as they did, but for the most part we found that the labels explain what DIR was expecting.

We worked within a few constraints during the pilot process. First, time: we only had six months to explore data from six agencies. Second, data sensitivity: we could not connect directly to their networks, or use any non-publicly available data. Third, platform: not all of Microsoft’s cloud offerings are available in the government cloud yet, so we had to run this pilot in the enterprise cloud. That meant it would be difficult to enable the pilot participants to connect to the environment with their work credentials.

Process

To execute this pilot, Catapult used our Agile Analytics for Government process.

image

This process is normally used to deliver one project for one agency or organization, but we adapted it to work for this pilot process. Key to our success were the jumpstart meetings. Typically a jumpstart focuses on the client’s business issues, their data, their constraints, their goals. This was no different, but we had couple of twists: almost everyone from the agencies was new to the Microsoft Cortana Intelligence platform, including PowerBI, Azure Data Lake, AzureML, Data Factory, etc… So, part of the jumpstart meetings entailed a hands-on crash course in the platform. This was necessary to help them understand the possibilities of a pure software-as-a-service business analytics and reporting platform. The other twist was that while the agencies brought real data that was really challenging them, we weren’t going to be able to deliver production solutions with this pilot. So, we had to focus on providing them some temporary insight while using that exercise to explore a capability for the greater purposes of the pilot.

After the six jumpstart meetings, we kicked off the first of ten two week sprints. Each of these sprints was preceded by a planning meeting where Nick the PM worked with agency stakeholders to take the jumpstart notes and consolidate them into agile epics and backlog items with acceptance criteria. During the kickoff meetings everyone collaborated to clarify requirements and criteria, prioritize, and assign roles. The stakeholders were given complete access to the platform and invited to work with us as much as they would like. Some agency personnel took advantage of the platform availability throughout the pilot and continued gaining insights from their data long after their sprint was over!

Coming in Part 2:

Infrastructure

Technology

Lessons Learned

Leave a Reply