Profile of a Campus Project
A research project to profile the University of Illinois at Urbana-Champaign campus.
Project Methodology
The Profile of a Campus Project is a research experiment to examine ways in which an academic institution can be profiled through a variety of techniques and data sources. Many of the techniques and data sources employed in this project have specific biases or nuanced interpretations that must be taken into account when using this data. For these reasons, these results should NOT BE CONSIDERED ACTIONABLE DATA, but rather suggestions of trends in the data which should be explored further.
Most sections do not employ name normalization for people or organizations, which can affect some rankings. This was done intentionally in this initial release to allow for further exploration of issues surrounding field normalization in data unification, which is one of the areas of my doctoral research.
THIS SECTION IS STILL BEING FINALIZED, GREATER DETAIL WILL BE PROVIDED HERE SHORTLY
The sections below provide greater detail on the methodology employed in each section:
Insights on Units
There are several notes on the main page of this section detailing the source of appointment data, which influences all sections using this data, such as space dispersion, correlating departments, etc. It is important to read the description for the "Multiple Appointments on Campus" as several of the fields are nuanced. Keep in mind that all locations used are the mailing addresses of those staff, rather than their office addresses, but in most cases those should be the same building. Users should pay special attention to the IBHE-Assigned Education Line for units that award degrees. This is a "best fit" that attempts to represent the department as best as possible through a single line, but many departments may award under multiple lines, so may not be completely representative of the total awards of a given department. In the Emphasis on Education there is a section called "UIUC By Year Code List" that lists all lines under which UIUC awarded each year, which can be consulted to determine what other related lines UIUC awarded under in a given year.
Focus on Engineering
No normalization of person, conference, journal, etc, names was done on this data. Therefore, "John Smith" and "Smith, John" will be treated a two separate persons. In most cases the majority of a given data field will follow a particular format (such as "Smith, John"), meaning that the lack of normalization should not affect results too significantly, but these issues should be kept in mind.
Emphasis on Education
Users of this data relying on the cohort comparisons should familiarize themselves with Carnegie Classifications and the list of institutions in UIUC's size cohort to determine any possible biases or assumptions with cohort computations. The Percent Minority graphs consider Asians to be a minority, which not all studies may consider to be a minority race. Especially note that the Line Group graphs use a non-standard grouping schema which will not match other studies, but which was used in this study as a experimental alternative.
Quick Graphs
These graphs are based on an all-campus median, which makes it possible to compare units from throughout campus, but makes the data less useful for examining a department's absolute performance. Also, certain categories (such as % College/School) are not meaningful for some units (such as GSLIS) due to the unit being its own School.
Inside Campus
While the strict numeric rankings of each unit along each indicator in each year are accurate, many of these data fields are EXTREMELY nuanced, and users of this data MUST consult the glossary linked on the main page of that section for a description of each indicator and how it is computed. Users of this section should ideally be familiar with DMI profile data and understand the context of each variable, and use this ranking as a way to help identify campus-wide trends. It is also important to keep in mind that medians are computed on a campus-wide basis and so are influenced by all units. Please read the introductory text on the main page of this section, as it describes some of the nuances of this data. In particular, ranks should not be used as absolute measures of a unit's performance in a particular indicator, as the values of other units above or below may be affected by specific nuances of the data. Uses of this data should focus on "big picture" analyses such as examining which departments on campus are ranked highest in terms of the percentage of their expenditures that come from state dollars/tuition, as those departments are the most vulnerable to state budget cuts.
Campus Concept Network
There is a great deal of discussion of the methodology used for this section on its main page. Specific points of bias on this page are the fact that only the top 100 most popular pages for each unit were considered and a naive maximal NP extractor was used, which affects the types of concepts and names identified by the system. The ranking of concepts was done by the total number of pages they appeared in, eliminating frequency within those pages, which is another source of possible bias.