Last week, I had the pleasure of representing my company, Infogix, at the Gartner Data & Analytics Summit in Grapevine, Texas. It was a great opportunity to talk data with people from around the country and across multiple industries, but another highlight for me was the Gartner Business Intelligence (BI) Bake Off. Now in its fourth year, this annual competition gives select BI vendors the chance to showcase their solutions to demonstrate how data and analytics can be leveraged for social good. We are producing data at a greater rate than at any other point in history, and we have the unprecedented ability to turn that data into tangible, actionable insights and information that can help solve some of the biggest issues plaguing society today.
This year’s bake off theme was “Doing Good with Data: The Opioid Crisis.” The data sets used in the bake off consisted of data from the Centers for Medicare & Medicaid Services (CMS), the Centers for Disease Control (CDC), and national census data. The participating vendors were asked to identify trends and intelligence based on that data. The vendors chosen for this exercise were among the industry leaders in BI. Non-BI vendors, such as Infogix, were given the same data and asked to put their spin on the scenario. It was a great opportunity to demonstrate how, with the right tools, raw data can be harnessed to uncover answers to a pressing social challenge metastasizing across the US.
BI tools provide great visualizations and insight into data and can drive business decisions; however, fundamental questions remain. Is the data accurate? Is this the right set of data? Where did the data originate from? Is the data current? Is the data useful? What other insights can be garnered from this data? These questions must first be answered in order to trust the intelligence produced by the BI solutions. Data must be validated, analyzed, and governed across the entire information supply chain. For the bake off challenge, after receiving the data, my colleagues and I set out to answer those questions using the Infogix Data3Sixty® platform.
The key to quality insights lies in quality data; if data sets are faulty or inaccurate, then no matter how powerful your BI or analytics tools may be, the truism of “garbage in, garbage out” will apply. To answer our first question, “is the data accurate?,” our first step was to create profiles of the data. These profiles measured various quality dimensions of the data: the number of null, empty or unique values, value fragmentation, and pattern fragmentation. Once incoming data meets initial quality standards to ensure its accuracy, completeness, and consistency, it must be tracked and validated as it moves across the hops from application to application, a process also known as integrity checks. An integrity check measures the attrition rate of the data, and any records that are changed, dropped, or added contribute to this attrition rate. Once the integrity checks are completed and we have a passing grade, the data is now deemed accurate and ready for use.
But data accuracy is just one of the crucial questions that must be answered to extract reliable intelligence from data. The other essential questions, (Is this the right set of data? Where did the data originate from? Is the data current? Is the data useful?) are answered through a strong enterprise data governance program. Our many discussions at the conference revealed that while most organizations have some governance in place, their solutions are overwhelmingly technical and only used by IT. In order for a governance program to be affective, it must bridge the IT to business divide and involve those business users closest to issues and tasked with making critical business decisions.
To build an effective data governance framework, using the data provided, we created a glossary of the business terms and defined them, as well as identified the synonyms for those terms. Ownership of the terms was then assigned to data stewards. These data definitions and synonyms enable users to choose the correct set of data for their analysis, and assigned stewards clearly delineate data ownership and designate responsibility for certifying and making changes to the definitions. Two other valuable governance tools that were used were an impact analysis map and business data lineage. The impact analysis allows users to see were the business term is used (e.g. applications, processes, policies, reports, etc.). Business data lineage allows users to see where data originated and the path it takes through the organization at a higher level, while technical data lineage traces data’s path at a deeper level. Now we know we have accurate data from the validation and the right data from governance. It is now time to begin our analysis.
BI tools offer great insights into the data, but we wanted to take a deeper look by using analytic models. For this exercise, we used a segmentation model to find correlations between overdoses and the amount spent on opioids. We also looked at which practices were prescribing opioids outside of the statistical norm. The results were then visualized in dashboards for analysis, and while some results were unsurprising given the current opioid crisis, some of the analysis outliers were eye-opening and raised additional questions that perhaps need to be explored. These issues could serve providers, regulators, and others who are on the front lines trying to combat the ongoing epidemic. If you would like additional information on our analysis, which was based on provider type and state, feel free to contact us.
The Gartner bake off turned out to be a valuable exercise on a number of levels. This exercise reinforced the power of data to solve some of society’s greatest challenges, and the ability of analytics to put answers that would have previously seemed out of reach right at our fingertips. But “doing good” with data must begin, quite literally, with “good data.” The keys to driving intelligent decisions through data is ensuring its accuracy and making it understandable and accessible to all users. The Gartner bake off was just one example of how accurate data can be used to solve a specific challenge. Of course there are many, many others.
To learn more about the platform used to solve this specific challenge and applying that level of scrutiny to your data to ensure smarter business decisions, check out the data sheet below.
For a deeper dive into this topic, visit our resource center. Here you will find a broad selection of content that represents the compiled wisdom, experience, and advice of our seasoned data experts and thought leaders.