Using AI to Break Down Head Count

New Microsoft® Power BI Visual Dynamically Analyzes Categorical Data

Like fashion goes in and out of style, so do visual ways of exploring data. Several decades ago I worked with a visual called the Decomposition Tree from an upstart company Proclarity from Boise, Idaho to analyze multi-dimensional data. That company was subsequently acquired by Microsoft but the visual never lost popularity. It was the single highest requested visual presentation in the Microsoft Ideas Forum.

What makes this visual so compelling is it allows the user to browse through categories of data effortlessly. Most categories or dimensions do not have to be thought of as hierarchical. Take head count for example. IPEDS head count has three built in categories of data – student level, race, and gender. The decomposition (decomp) tree allows you to explore the different categories in sequence – the more categories, the more flexibility. The decomp tree provides a level of flexibility that citizen data scientists love. The three minute video illustrates the power of this tool.

The artificial intelligence features of the decomp tree are Microsoft’s attempt to provide built-in intelligence to the visual. We might think of these tools as consumer artificial intelligence. The AI component of the visual allows you to determine which categories have the most or least influence on the visual.

The Insight for Higher Education (IFHE) dataset provides built in capability to take advantage of the decomp tree. For example, there are over 100 dimensional attributes that describe higher education institutions in the dataset. In the video example below, we show several ways of analyzing head count using the decomp tree.

We have created a public visual below that allows you to play with the consumer features of the decomp tree. Simply select the institution and click on the nodes. You can also close categories and play with different category paths such as race, then gender, then level.

About Public Insight

Public Insight transforms the islands of public data into actionable insights with analytic datasets and Microsoft® Power BI to better understand your industry and market. Insight for Higher Education is an easy-to-use executive planning and analysis application for benchmarking higher education institutions across a wide range of topical areas such as enrollment, graduation rates, degrees awarded, and staffing. Sign up for a free Factbook Interactive to explore how easy it is to navigate through public data.

Share This Story

Similar Posts

  • Student Earnings Experience a Sluggish Recovery

    College students are earning 8% less in 2013 than in 2008 per the most recent data published by the US Department of Education College Scorecard database. Students earned $31,382 in 2013 compared to $34,119 in 2008. The earnings were ostensibly flat from 2012 earnings of $31,528.   The College Scorecard…

  • Will Voter Unpredictability Continue in 2018?

    Much has been written about the increased polarization of America and its collective impact on communities. There were many surprises in the 2016 election and people are bracing for a potentially tumultuous 2018 mid-term election. Ohio, where I live, is one of the swing states. The average county in Ohio…

  • New Violence Against Women Data on Campuses

    A slug of new data has recently been released covering campus statistics on violence against women. This data was mandated by the Violence Against Women Reauthorization Act of 2013 (VAWA) and is added to numerous other statistics required under the Campus Crime Statistics Act, or Clery Act. College institutions must…

  • Hot Growth Industries

    The building boom is officially back after a long hiatus. The number of construction employees increased 5% in 2016 according to the most recent County Business Patterns released by the U.S. Census.  The sector however with the most number of new establishments is the Information sector (NAICS 51) which increased 5%….

  • Nursing Home Coverage of the Elderly

    While long-term care services more than the elderly, approximately 63% of residents are 65 or older. According to the Administration on Aging, population 65 years or older will increase from 14.1% of the population to 21.7% of the population by 2040. As the population ages over the course of the…