Airline Passenger Satisfaction Analysis

---

Introduction

Oceny satysfakcji pasażerów pokazują nie tylko subiektywne wrażenia z podróży, ale także jakość procesów obsługi i przebiegu lotu. Ten projekt powstał jako ćwiczenie analityczne i miał na celu uporządkowaną analizę ankiet pasażerskich oraz identyfikację wzorców, które wpływają na ogólną ocenę lotu.

The dashboard was designed to present satisfaction not as a single average, but as a set of segment-dependent relationships. The analysis explores how ratings vary by travel class, customer type, purpose of travel, demographics, and delays. The focus was on creating a clear and restrained visual design that supports interpretation and comparison across groups, rather than reducing the results to a single metric.

The project covered the full data preparation process and the development of an interactive analytical tool in Excel.

Data

• Scope of analysis: It includes 19,652 survey responses.
• Data source: The data comes from the training dataset “Sky Is The Limit”, provided as part of the course. The materials did not include a publicly available primary data source.
• The data was provided in a tabular format and included, among others:
  • Passenger demographic information
  • Customer type
  • Type and purpose of travel
  • Travel class
  • Flight distance
  • Delays
  • oceny poszczególnych elementów podróży, takich jak komfort, obsługa, boarding czy WiFi.

Process

1) Data preparation and modeling in Excel

• The analysis was conducted in Excel using worksheets that document each stage of the process.
• The key steps included cleaning and standardizing the survey data, creating derived variables, restructuring rating fields into a format suitable for factor analysis, and building supporting tables for Top 3 and Bottom 3 comparisons. Maintaining a consistent aggregation logic was essential so that all visualizations were based on the same metric definitions.
• The visualizations were based on pivot tables designed to enable dynamic filtering and comparison across segments.

2) Visualization and interaction in the dashboard

• The dashboard was designed as an exploratory tool that supports interpretation and comparison, rather than as a static report.
• The structure includes an overview page showing overall satisfaction and demographics, as well as a factor analysis page presenting ratings of travel components and comparisons between segments and the overall score.
• I used KPI tiles, column and line charts, a donut chart, and a radar chart that shows the profile of a selected segment against the overall average.
• The results can be filtered using slicers. In addition, I implemented a VBA-based button to reset all filters.
• The layout and color scheme were selected with a focus on readability, comparability, and visual consistency.

Results

• Analiza pokazała niski ogólny poziom satysfakcji pasażerów oraz silne różnice między segmentami. Widoczny był duży wpływ klasy podróży na poziom zadowolenia, a także istotny spadek satysfakcji wraz ze wzrostem opóźnień. Różnice były również wyraźne między podróżami służbowymi i prywatnymi.
• The lowest-rated areas were Wi-Fi, the booking process, and punctuality.
• Dashboard umożliwia szybkie porównanie segmentów oraz identyfikację elementów podróży, które najsilniej obniżają ocenę lotu.

Reflections

• The biggest challenge was not the visualization itself, but maintaining a consistent analytical logic across multiple segmentation dimensions. It was crucial to define metrics consistently, ensure comparability of results across charts, and strike a balance between detail and readability.
• Projekt ma charakter demonstracyjny, ale pokazuje sposób myślenia analitycznego oraz podejście do projektowania dashboardu jako narzędzia wspierającego interpretację i decyzje.