Decoding Insights from Airline Reviews Using Sheet Assistant

In an era where air travel has become vital to global connectivity, recent challenges have spotlighted the critical importance of passenger experiences. From sudden cancellations to overcrowded flights, the aviation industry also grapples with many issues that underscore the need for enhanced services and technologies.

Consequently, as passengers embark on journeys that span continents, a slew of questions arises: How do passengers genuinely feel about their air travel experiences? What elements elevate their satisfaction, and which aspects contribute to their discontent?

This article embarks on a compelling journey, merging the world of airline reviews with the transformative capabilities of Gigasheet's new AI powered Sheet Assistant. From performing sentiment analysis using AI to predicting trends, our latest feature transcends analytic boundaries and ushers in a new era of enhanced understanding and optimization, helping companies achieve data-driven excellence.

Let’s dive in.

Enter Sheet Assistant

Gigasheet now includes an AI-powered assistant to help you easily understand your data. This cutting-edge tool intelligently analyzes your data based on prompts. Simply ask your questions, and the Sheet Assistant will filter, group, and aggregate your data to present you with meaningful insights.

At Gigasheet, we believe that Sheet Assistant can revolutionize the way large datasets are analyzed, allowing researchers to easily navigate complex data landscapes, uncovering trends and correlations while promoting our goal of democratizing big data.

But before we jump to it, let’s familiarize ourselves with our dataset.

Understanding the Airline Review dataset

The "SkyRatings" dataset used in this post is a comprehensive collection of passenger reviews and ratings covering various aspects of air travel scraped from https://www.airlinequality.com. This website allows travelers to share reviews and ratings about different airlines, including destinations and experiences, where consumers freely share their thoughts on flights, amenities, customer service, and more.

Here are some additional details about the dataset and some possible uses:

‍Attributes: Each row in the dataset includes features that capture different aspects of the air travel experience—these include airline names, aircraft types, seating comfort, and amenities such as in-flight entertainment.

‍Temporal Aspect: The dataset includes reviews spanning a few years, offering insights into how airlines' services have evolved and how passenger sentiments have changed.

‍Sentiment Analysis Opportunities: Given the presence of review content and associated ratings, sentiment analysis can extrapolate the overall emotional tone of the reviews—this can unveil broader satisfaction, frustration, or neutrality sentiments among passengers.

Pattern Identification: Different techniques can uncover patterns and trends in a dataset. For example, you could identify recurring issues, frequently mentioned amenities, or aspects that consistently receive positive or negative feedback.

Customer Experience Enhancement: The dataset's insights can guide airlines in refining their services. Airlines can focus on enhancing specific areas by understanding what aspects contribute to positive or negative passenger experiences.

Airline Experiences Unveiled

Now it’s time to put Sheet Assistant to the test by unleashing its analytical potential on our dataset to perform sentiment analysis. One possible direction is to leverage the generated examples, or recommended prompts, once Sheet Assistance has had the time to traverse every row.

For instance, let’s click on the first example: “Show me the rows where the Overall_Rating is greater than or equal to 4.”

‍

Interesting stuff! Sheet Assistant has automatically filtered the relevant column (Overall_Rating) for us. It even presents us with an additional set of columns we might be interested in.

Next, let’s focus on “Verified” reviews only—this adds a layer of trust for analysts, as they can have confidence that the shared experiences are from legitimate sources. A simple filter can get us there:

‍

With our dataset containing verified reviews only, let’s engage in pattern identification by asking Sheet Assistant to help us with the following question: “Do reviews from business travelers differ significantly from those of leisure travelers?”

Let’s begin by asking Sheet Assistant to group reviews by traveler type:

‍“What kind of travelers are there?”

Sheet Assistant “knows” exactly how to group these records based on the “Type of Traveler” column; in fact, it even tells you the steps involved:

‍

The results are as follows:

‍

In filtering terms, this is akin to:

‍

Using numeric aggregations (average) on these four groups, we can reflect the overall ratings associated with all aspects of the airline experience:

‍

Now, you’d be asking: can’t Sheet Assistant do everything in one step? Of course! Just tell it to:

‍“Divide the type of travelers into groups for verified reviews only, and show me the average overall ratings for each group.”

And, voilà!

These ratings reflect distinct perceptions among different traveler groups. For instance, business travelers give a higher overall rating, indicating a more positive view of the airline's services. In comparison, solo leisure travelers exhibit a slightly lower average rating, suggesting that they might have encountered some aspects that affected their experience.

However, the ratings provided by couples and family leisure travelers are notably lower, meaning that these groups might have experienced more significant challenges or issues during their trips.

A second scenario may entail Value for Money—analyzing whether passengers felt they received good value for the ticket price. In particular, segmenting the data by different traveler types (business, solo leisure) and analyzing their “Value for Money” ratings can reveal if certain groups have different expectations or perceptions about the cost-effectiveness of their travel.

Luckily, we have a column in the dataset that accounts for this. We’ll begin by asking Sheet Assistant:

‍“For travelers with verified reviews only, group traveler types by average scores based on value for money and only show me this column.”

Resulting in:

‍

Noticeably, the higher values of 2.476 and 2.411 for "Solo Leisure" and "Business" travelers indicate that these groups, on average, perceive the airline experience slightly more favorably in terms of value for money. These scores closely match the “Overall Rating” ones, suggesting that solo leisure and business travelers might find the airline experience relatively more pleasing than the other two groups.

Beyond the above insights, these data can unveil (further) hidden correlations between aspects such as customer service and entertainment preferences. By delving deeper, these data hold the potential to guide airlines in refining service offerings, optimizing passenger experiences, and ultimately elevating the journey from boarding to disembarkation.

Final Thoughts on Performing Sentiment Analysis Using AI

At Gigasheet, our commitment lies in providing innovative solutions that empower businesses and researchers to explore, analyze, and extract value from their datasets with unprecedented efficiency and simplicity. With this ethos at the core of our mission, we continuously strive to enhance how data is understood, visualized, and utilized, propelling organizations toward evidence-based practices.

Sheet Assistant has democratized big data analytics by making it accessible to anyone with spreadsheet skills. We continuously add new features and functionality to Sheet Assistant, so stay tuned for more! And sign up for free to try it yourself.

‍