This is a group practicum which requires collaboration in groups of three to four students. Students are automatically assigned to groups via Canvas and are expected to fully collaborate and submit the same work. Peer evaluations will impact final grades.
The practicum focuses on the National Transportation Safety Board (NTSB) Aviation Accident & Incident Dataset. The NTSB publishes data on select aviation accidents and incidents within the United States, its territories, and international waters. The dataset is updated daily and tracks the status of each event through various investigation stages, from preliminary reports to the final probable cause.
The goal is to perform a comprehensive exploratory data analysis (EDA) and present actionable insights that inform:
- New pilot training requirements
- Aircraft inspection protocols
- Weather-related flight restrictions
- Insurance regulations
- Public safety communications
As the NTSB Chair emphasized: "We need to understand not just what happened, but why it happened, where it's happening, and most importantly - what we should do about it. Every visualization should tell us something actionable."
- Source files: Download source files
- Data dictionary: Download data dictionary
-
Create an executive dashboard showing the current state of aviation safety:
- Line chart: accidents vs. incidents over the most recent 20 years
- Proportion chart: fatal vs. non-fatal events over time
- KPI/gauge: current "safety score" (define metric and calculation factors)
-
Insight Goal: Determine if aviation safety has improved or stagnated and identify concerning trends.
- Multi-dimensional analysis using 4–5 visualizations.
- Analyze combinations of risk factors (e.g., weather, aircraft type, flight purpose).
- Investigate hidden risk patterns and risk evolution over time.
- Insight Goal: Identify combinations of factors with highest risk and emerging patterns.
- Analyze which flight phases most accidents/incidents occur.
- Break down events by another dimension to draw insights for each phase.
- Analyze aircraft models with disproportionately high accidents/fatalities.
- Visualize concentration across flight phases.
- Evaluate safety trends over time. (Focus on 5 manufacturers)
- Use number of engines and engine type as complexity indicators.
- Visualize patterns for more complex aircraft.
- Compare accident severity across at least two aircraft categories.
- Map accident density by state/region.
- Identify persistent danger zones.
- Calendar heatmap for accident frequency.
- Identify seasonal patterns and periods needing extra safety measures.
- Conduct a comparative analysis of choice with at least 5 visualizations.
- Provide complete justification and storytelling using the dataset.
- Summarize overall analysis.
- Include 3 key findings, 2 critical recommendations, and an aviation safety assessment.