Airline Industry
Customer Satisfaction Analysis
Project Snapshot
The Project Objectives
Project: Data Analysis Class Assignment
Client: Airline Industry
Timeline: 6 weeks
Team: Three team members, varying ages & skillsets
Role: Data scientist, slide designer
Skills: Data Analysis, Storytelling, Graphic Design, Solution Development, Customer Segmentation
Methodologies: K-Means Clustering, R-Studio, R Coding Language, Large Language Models
-
Discover meaningful customer segments through cluster analysis using R
-
Determine service factors with high influence using multiple linear regression
-
Compare satisfaction ratings among clusters to identify areas for improvement
-
Draw insights and make marketing recommendations
Clustering Methodology
Goal: Identify traveler segments based on demographic and flight characteristics.
Variables Used: Age, Type of Travel, Class, Flight Distance
Data Preparation:
-
Cleaned missing data
-
Encoded categorical variables numerically
-
Scaled numeric variables for equal weighting
Method: K-Means Clustering (3 clusters, nstart = 50)
Cluster evaluation: Based on silhouette, satisfaction scores, and interpretability
Cluster profiling: Used centroids to describe each segment
Research Methodology
1.
Dataset Cleaning & Familiarity
Our group began with cleaning the data before running analysis (as seen above in data preparation) and became familiar with the variable available to cluster based off of. We then chose the variables listed above to cluster in R-Studio based off of.
2.
K-Means Clustering (n=3)
Details of clustering choices can be seen in Clustering Methodology section above.
3.
Regression & Correlation Analysis
Details of analysis choices can be seen in Regression & Correlation Analysis methodology section below.
4.
Insights & Managerial Recommendations
Taking both the clustering of consumer base, regression analysis, and correlation analysis into account, our team found multiple recommendations and managerial insights for airplane companies. The data informed decisions can be found in the slideshow above or managerial recommendations below.
Regression & Correlation Analysis Methodology
Objective of Analysis: Identify which controllable service and operational factors most influence the KPI of SatisfactionScore
Descriptive and Correlation Analysis Process:
-
Chose independent variables of OnboardService, Cleanliness, FoodDrink, BaggageHandling, LoyaltyProgram, InflightWifi, SeatComfort, EaseOfCheckIn as they ranked lower on the descriptive analysis - meaning they had on average lower ratings - or would have higher ROI for the cost by the managerial
-
Correlation analysis highest correlation variables being OnboardService, BaggageHandling, LoyaltyProgram, SeatComfort, and InflightWifi, meaning
Project Managerial Recommendations
1. Address issue of low seat comfort for the consumer group of Practical Business Flyers
-
-
Has greatest impact on satisfaction and increasing seat flexibility maintains high ROI w low cost
-
2. Implement RFID to solve issue of low baggage handling experience for Practical Business Flyers
-
-
Has large impact on satisfaction
-
Have case of Delta Airlines to prove its viability to pitch to those in greater managerial positions
-
Relatively high ROI with low upfront capital cost due to being technology-based solution
-
3. Lastly, work on improving onboard service for Economy class, specifically targeting Business flyers
-
-
Slowly introduce technology integration and increased training to staff
-
Has slightly lower ROI and upfront cost than other solutions, but has long-term impacts on consumer relationships, validating cost overtime
-
Project Showcase & Reflection
This project taught me the importance of leveraging hard skills, such as R-Studio coding, K-Means Cluster, Multi-variable regression & correlation analysis, alongside Generative AI to successfully write replicable code to produce a safe, clean, and truthful data output. In this project, I was able to spend time learning a new coding language and benfitted heavily from relying on supportive material, at home research, and professor aid.