THANK YOU FOR SUBSCRIBING
When Governor Rauner tasked IT leaders in Illinois to embark upon IT Transformation, it quickly became apparent that data sharing would be a critical component. Work began on an Enterprise Memorandum of Understanding (eMoU) that provided the vehicle for thirteen of our largest Health and Human Service agencies to seamlessly share data. This foundational event set in motion the activities that put Illinois state government on a path to becoming a data-driven enterprise. To realize the vision, a Statewide Data Practice (SDP) with a vibrant strategy and a long-term view was established to assimilate and proliferate the art of transforming data into actionable insights.
1. The Methodology
It is important to recognize that the concept of creating value from data has been around since time immemorial. The method of converting data into useful insights is a scientific and disciplined practice. The practice needs to be communicated and taught across the organization in order to propagate the ongoing revolution. The volume of data that is being created is growing exponentially while the storage and computing costs are dropping to create this opportune and revolutionary moment. In order to realize the full potential of this transformation, newer technologies need to be phased in slowly and steadily. Any organization embarking upon the same journey needs to craft at least a four-year plan that is fully endorsed by executive leadership and communicated widely for support from across the organization. The long-term plan should include clear short-term milestones that are necessary to demonstrate the value of a wide variety of solutions.
"The method of converting data into useful insights is a scientific and disciplined practice that has to be communicated and taught across the organization in order to propagate the ongoing revolution”, Krishna Iyer"
2. Take an Enterprise View
The true value of shared data occurs at the intersection of two or more distinct entities, or agencies in the public sector, led by separate and distinct executives. The Chief Data Officer (CDO) leading the practice needs to possess the leadership skills necessary to unify business owners and leaders, as well as agency leadership including Chief Information Officers and Directors. The CDO serves in an advisory role on the project team as technology solutions are considered to be run on the enterprise platforms. The CDO maintains the eMoU, establishes an enterprise vision for statewide data architecture, maintains appropriate technology and tool stacks, promotes data literacy, creates standards, and enables data governance. It is the expectation that the CDO will gauge the situation at every agency and devise a solution that addresses that agency’s basic needs before encouraging more advanced solutions.
3. Create Data Analytics Practices at Agencies
The agencies play a role distinct from that of the SDP, yet as integral a part of the enterprise plan. Subject matter experts (SMEs) should reside within the agencies to address those questions particular to their agency while acting as the internal marketing arm to promote data sharing and best practices. The SDP will act in a supporting role from the enterprise perspective. Staff from the SDP will train the agencies in the methodology, beginning with exploratory studies and ending with a robust machine learning model that can be deployed in production. Eventually, the agency data practices will become less reliant on the SDP for day to day support and the SDP can return to more of an enterprise and strategy support model.
4. The Functions
The readiness of agencies to absorb analytics falls across a wide spectrum. It is important to assess an agency’s needs and propose a solution that is suitable for their individual circumstance. The support functions derived from a data practice are of two varieties:
5. Business Intelligence
This support is the most traditional and basic of all data practice functions. The primary objective is to support decision-making by explaining information provided by gathered data in a multi-dimensional and easy to digest interactive visualization. This function serves as a pre-requisite for Advanced Analytics by requiring that essential data sets are in order, by verifying data integrity and by improving data accuracy. Commonly available tools can provide direct access to curated data sets, enabling live reporting and dash boarding making the process fast and reliable for decision making.
6. Advanced Analytics
As the agency matures in their data practice and begins to appreciate the value of business intelligence, they can take advantage of more advanced tools like supervised and unsupervised machine learning methods. These advanced tools can be embedded within the process, are amenable to automation and rely on availability of computing and storage. Common applications for these tools are: population clustering and classification, student scoring, patterns of unemployment, recidivism tendencies, to name a few.
3. The Platform
The effectiveness of the SDP in delivering value to the enterprise is highly contingent on the tools that are available for solving a given use case. A platform, that can maintain a tool stack for mastering and sharing data, conducting scientific explorations, developing models and with most up to date algorithms, is absolutely essential. The models can be made available through the platform for agencies to digest, score or classify data. Another benefit of an enterprise platform is the capability of measuring the return on investment (ROI) of data science projects beginning with conception through implementation. It will also be capable of handling a variety of needs with a range of tools across the enterprise. As the models mature, a reliable pipeline of data is created. The models tell a data story that can be recited across the enterprises, showcasing the capabilities and benefits of data analytics. The platform will also support other functions of SDP like data governance, open data portal for transparency, etc.
4. The People
The State Data Practice (SDP), tasked with propelling this important agenda should be staffed with the talent and experience to meet the demands of the projects. The SDP should have skill-sets that match the stages of the data science life-cycle with enough agility to work with a wide variety of agency challenges. Recommended roles for a practice are: Data Scientists, Data Engineers and Data Architects. Their skill-sets should span a wide range and include the ability to create user-friendly, insight rich interactive dashboards, the knowledge of advanced algorithms and most importantly, the resources to deliver a center of excellence rich in the abilities to teach, educate and coach the budding data scientists and those aspiring to transition into this field. Small successes by the team should be celebrated along the way.
In summary, executing on this strategy will go a long way toward putting the State of Illinois truly on the path of digitization and will enable the drive toward adopting other modern platforms, blockchain and artificial intelligence. Most importantly, following the strategy will provide citizens with the services they truly deserve.