Define comparable groups
Classify rest opportunities, crew role, nap status, and flight timepoints before making a comparison.
CASE STUDY 06 / MODELING
Inflight rest and pre-flight nap analysis
Comparing rest patterns while keeping subgroup size and uncertainty visible.
FOUR ROUTE ANALYSIS SCOPE
Illustrative visual; no research data shown
THE QUESTION
How do fatigue, sleepiness, sleep, and cognitive measures vary across rest schedules and pre-flight nap groups?
THE SETTING
The analysis classifies first, middle, second, and split-break patterns across four routes, separating flying and relief crew and examining multiple measurement timepoints.
THE STORY IN PLAIN LANGUAGE
A difference needs context before it becomes an explanation.
Operational groups differ in rest opportunities, crew roles, and measurement timing. The analysis moves from those definitions to subgroup availability, visual comparisons, and effect-size estimates so the size and uncertainty of a pattern can be understood together.
Technology behind this workflow 11 tools & formats
The stack follows the analytical task: prepare the data, fit or apply the method, check the result, and communicate the output.
See how the libraries connect to the wider portfolio →THE PROJECT, STEP BY STEP
See how the question became a reviewable result.
06 / MODELINGVISUAL WALKTHROUGH
STEP 01 / 03Define comparable groups
Define subgroupsSubgroup countsContrasts + intervals
Four route analysis scopeSTEP 02 / 03Show the denominator
Define subgroupsSubgroup countsContrasts + intervals
Four route analysis scopeSTEP 03 / 03Quantify differences carefully
Define subgroupsSubgroup countsContrasts + intervals
Four route analysis scopeIllustrative process diagram
Show the denominator
Produce subgroup counts and complete-case availability tables so sparse comparisons are visible.
Quantify differences carefully
Use timepoint plots, Welch comparisons, confidence intervals, and effect-size estimates for observational patterns.
THE TECHNICAL CASE
Decisions behind the analysis.
Operational subgroup comparisons need careful definitions and visible sample availability. The work connects rest schedules and pre-flight nap patterns with fatigue, sleepiness, sleep, and cognitive-performance summaries.
- Unit of analysis
- A trip/crew subgroup measured at specified flight or rest timepoints.
- Inputs
- Rest opportunities, crew role, nap status, flight phase, performance measures, fatigue/sleepiness scales, and sleep summaries.
- Deliverables
- Group counts, complete-case tables, timepoint plots, Welch comparisons, confidence intervals, and effect-size summaries.
Define comparable opportunities
The reports distinguish first, middle, second, and split-break patterns and separate flying from relief crew across four routes. The same label must represent comparable timing before a pooled contrast is interpretable.
Account for available measurements
Subgroup counts and complete-case availability tables clarify which measurements enter each comparison. Repeated timepoints can have different denominators even within the same operational group.
Quantify magnitude and uncertainty
The implemented notebooks use Welch comparisons, confidence intervals, and effect-size estimates including Cohen’s d. Welch’s procedure relaxes equal-variance assumptions for a two-group mean comparison but does not solve confounding or repeated-observation dependence.
Keep timepoint context
Plots summarize outcomes at flight and rest timepoints. A start-break measurement and a top-of-descent measurement answer different questions, so the subset used for a test must match the label in the report.
METHOD IN FOCUS
Observed group difference = mean(group A) − mean(group B)
A difference is an observational association unless the design supports a causal interpretation. Its uncertainty and practical scale matter alongside its p-value.
EXPLORE THE METHOD
Read an effect estimate together with its interval
Synthetic differences and 95% intervals illustrate precision. They are not findings about actual nap or rest groups.
QUALITY OF THE EVIDENCE
Checks that change the conclusion.
- Subgroup definitions
Check crew role, break type, route, nap status, and timepoint before comparing records.
- Sample availability
Display usable outcome counts rather than implying all trip records contribute to every endpoint.
- Multiplicity and exploration
Treat a broad set of subgroup/timepoint comparisons as exploratory unless a confirmatory plan specifies the contrasts.
- Causal limits
Do not attribute improved performance to naps or claim an optimized operational schedule from observational comparisons alone.
FOR THE TECHNICAL CONVERSATION
Where I would take the discussion next.
An interviewer can probe repeated-measures dependence, multiple comparisons, confounding by crew role or route, and how to present an effect size when subgroup estimates are imprecise.
WHAT IT DEMONSTRATES
From method to meaning.
The reports provide interpretable subgroup comparisons and visual summaries for research discussion.