Airline Passenger Satisfaction Survey Analysis
A survey is usually carried out on a selected sample of individuals to obtain data about a service, product, or process. This information is then analyzed to gain insights into the respondents’ patterns and behaviors toward the given service, product, or process. Most businesses across the world utilize surveys to sample the opinions of their customers or clients toward their products. Insights from these surveys help stakeholders make informed decisions that improve their businesses’ overall potential and performance.
Most times organizations make use of the Likert scale to scale responses or opinions from their customers. Depending on the nature of the responses, the scale could be 5 points or 7 points. The responses usually range from a high to low, or strongly agree to strongly disagree, etc. They also come with a midpoint for those opinions which are neutral.
For this project, I’ll be working on an airline passenger satisfaction survey from Kaggle. The dataset can be found here. I cleaned, analyzed, and visualized the data using Powerbi Desktop.
Data Cleaning
1) Changed the “ID” column data type from numeric to text
2) Replaced all null values in the “Arrival Delay” column with “0”
Data Transformation
- I duplicated the airline_passenger_satisfaction table but retained only the “ID” and 14 columns belonging to the customer survey responses.
2. I unpivoted the 14 columns showing the customer survey responses to generate 2 columns showing the attribute column and their corresponding values.
3. I added a custom column showing the Likert value of the responses:
5= “highest”, 4=”high”, 3=”neutral”, 2=”low”, 1=”lowest”, 0 = “not applicable”.
4) I created the following DAX measures to help analyze the data further:
Avg arrival delay, Avg flight distance, Avg departure delay, count all, high count, highest count, neutral count, low count, lowest count, not applicable,% high score, % highest score,% neutral score, % low score, % lowest score, %not applicable score, NPS(Net promoter score)which is the difference between the high and low scores [100 * (([% highest score] +[% high score])- ([% lowest score]+ [% low score]))], and the Rank NPS = RANKX(ALL(airline_passenger_feedback[Attribute]),[NPS],, DESC).
Data Visualization
- I used pie charts to visualize the distribution of the following: customer type, Type of travel, Class, and customer type.
2. I used a column chart to visualize the satisfaction rating of passengers.
3. I used a histogram to visualize the Age distribution of passengers.
4. I used a stacked column chart to visualize the survey responses. I placed the attribute column (survey questions) on the y-axis and the % scores on the x-axis. I started with the %not applicable score first then worked my way up to the %highest score.
5. I created a matrix showing the attribute column, and the NPS, and then cropped the chart to reveal only the NPS score. I formatted the matrix and included a background color to enhance the visual. I sorted the NPS in descending order and placed the matrix right next to my stacked column chart. I also inserted an arrow to show the direction of the increment.
DASHBOARD
INSIGHTS
The result of the survey showed that most passengers rated their in-flight and baggage handling services high at 36% and highest at 26%. They equally enjoyed their experience in flight such as the leg room, seat comfort, onboard service, check-in service, and inflight entertainment However, the reverse was the case with the ease of online booking where most of them rated their experiences low at 25% and lowest at 17% and inflight wifi service recording low at 25% and lowest at 17%. The airline will do well to improve its online booking process and inflight wifi service for its passengers to maintain their loyalty and increase patronage.
References and Resources
- ASQ(n.d). What is a Data Collection Survey? https://asq.org/quality-resources/survey.
- Typeform(n.d). What is a Likert scale survey, and what are they good for?https://www.typeform.com/surveys/likert-scale-questionnaires/
- https://youtu.be/n5yG7oetflc
