February 2026

Melton Scholarship Mini-Case

Overview

The Melton Scholarship mini-case provides students experience in conducting guided research. Student perform weekly tasks with mentors. Unlike typical homework, however, these tasks are much more open-ended, require substantial thought, take trial and error to solve and might not have one “correct” solution. Review the following information for typical mini-case expectations.

Background

Mini-Case Subject: Airplane Delays

Chicago O’Hare (ORD) is one of the nation’s biggest airports. In winter months, major weather events have the possibility of wreaking havoc and severely delaying or canceling flights.

Based on that scenario, Melton faculty mentors might ask the following questions: How big of an impact do major events have on the departure delay and cancellation rate of outgoing flights from ORD? How do delays that originate in O’Hare propagate across the country and influence delays elsewhere?

Task

As part of the Melton Scholar application, include a short response to those mentor-posed questions so the application committee can understand how you approach the problem: what data felt necessary to gather, what seemed worth measuring, which comparisons made sense and what visuals could help make things clearer? The focus isn’t on getting perfect results, but rather on the thinking behind the work and the choices made along the way.

In a real project, this would be an initial pass; sharing early findings with a mentor, then working through what they mean, what could be improved, and where to go next. Research rarely lands on the right answer the first time. It’s a back-and-forth process that includes reworking ideas, trying new approaches and building a clearer understanding step by step until a strong path forward takes shape.

Application Response Guidelines

  • No longer than three to four pages
  • Walk through the approach, not just the outcome, to show how the analysis took shape, including what was explored, what was measured and how conclusions were reached
  • It’s not necessary to cover everything mentioned, but the work should connect clearly to at least one of the mentor’s prompts
  • Include some numbers to support the story; the addition of a clear, well-designed visual that strengthens the takeaway provides more support
  • Use whichever tools fit the approach: Excel, R, Tableau, Python or others
  • Email Adam Petrie (apetrie@utk.edu) with questions
  • Submit a Word document or PDF to Adam Petrie after submitting the application no later than the date specified on the Melton Scholar website

Obtaining Data

The Bureau of Transportation keeps an online, queryable database of nearly all aspects of every U.S. flight, including delays. As of February 2026, the best way to obtain data is to visit: https://www.transtats.bts.gov/DL_SelectFields.aspx?gnoyr_VQ=FGJ&QO_fu146_anzr=

Select Geographic (to get a specific state if you necessary), Filter Year and Filter Period accordingly to get the necessary data. Users are restricted to one month at a time.

Reporting Carrier On-Time Performance (1987-present)

Select fields to download. For the purposes of this study, some selectable, applicable fields are:

  • FlightDate (or other temporal field)
  • Reporting_Airline (see lookup table) – name of airline (e.g., Delta, United)
  • Origin – name of airport that airline is departing (e.g., TYS [Knoxville] and ORD [Chicago])
  • Dest – name of airport to which the airline is arriving
  • CRSDepTime – scheduled departure time
  • DepTime – actual departure time
  • DepDelayMinutes – difference in minutes between scheduled and actual departure times
  • DepDel15 – 1 = delayed by 15 or more minutes, 0 otherwise
  • CRSArrTime – scheduled arrival time
  • ArrTime – actual arrival time
  • ArrDelayMinutes – difference in minutes between scheduled and actual arrival times
  • ArrDel15 – 1 = delayed by 15 or more minutes, 0 otherwise
  • Cancelled – 1 = yes flight was canceled, 0 otherwise; note the Code column as well
  • WeatherDelay – delay due to weather, in minutes

Additional Guidance

  • Big, open-ended problems usually start small by focusing on one piece of the picture, then building from there; starting with a narrower, more manageable question is often the best approach
  • COVID-19 disrupted flight patterns in a big way; focus on timeframes where that impact is minimal, either before March 2020 or after mid-2022.
  • The data source provides information by month; pulling multiple months may require a few manual downloads.
  • Identifying “major events” will take some digging; AI tools can help point to sources for historical weather data
  • Plan to compare something like average departure delays at ORD on major event days versus typical days; that means pulling in a few “normal” days as a baseline, so weather data will be useful more than once
  • December 4, 2016, stands out as the first major event of the season, heavy wet snow and the largest December snowfall at ORD in over a decade; that could be a strong starting point and a nearby date without notable weather could serve as a comparison