Researching Instagram Sentiment and providing an Analysis
How has users’ sentiment of Instagram App Store reviews changed via location?
- How does text sentiment differ from numerical ratings?
- What themes appear in positive vs. negative reviews?
- Which regions show higher or lower user satisfaction?
Understanding how user sentiment varies across geographic regions is essential for evaluating the global performance of digital platforms. As Instagram continues to expand its international user base, the App Store reviews left by users provide valuable insight into how experiences, expectations, and reactions differ across countries.
The raw data for this project was obtained by us from a publicly available dataset on the Kaggle platform (Kanchana1990, 2023). This dataset was collected by the author from real reviews on the Apple App Store's review pages and has been organized by the author into a structured CSV format. Since the original publisher has already completed the data collection and anonymization process, we started working directly by downloading this dataset and did not involve the web scraping step.
- 500 user reviews
- 11 variables
When cleaning the dataset to get the reviews, we had to:
- keep only required columns
- convert date format
- convert score to numeric
- merged title + text into one complete review field
- create time variables
Sentiment analysis helps us identify whether each review is positive, negative, or neutral, allowing us to understand users’ true feelings beyond star ratings.
This allows us to more accurately understand how users’ opinions of Instagram have changed across different years, app versions, and countries. It also helps identify patterns in what users like or dislike, providing deeper insights that ratings alone cannot show.
We use this method because star ratings alone do not fully reflect users’ true feelings. A user might give a low rating but write a positive comment, or give a high rating while expressing frustration in the text. By analyzing the written review itself, sentiment analysis helps us capture the actual tone and meaning behind each review.
We can see that positive and negative reviews are relatively close (196 negative and 185 positive). There is 119 reviews classified as neutral, indicating a mix of information or emotionally balanced comments.
Germany shows the highest level of dissatisfaction:
- Both the average star rating and sentiment score for Germany are significantly lower than all other countries.
- This suggests that German users may be experiencing more issues or expressing stronger frustration with the app during this period.
India and the United States are the most satisfied:
- India and the U.S. have the highest average ratings and the most positive sentiment scores among all regions.
- This indicates that users in these countries generally have a more favorable experience with the app.
There is a mismatch between ratings and sentiment in some countries:
- Canada (CA): The average rating is relatively low, but the sentiment score is close to neutral.
- United Kingdom (GB): The average rating is moderate, yet the sentiment score is slightly negative.
Positive Word Cloud:
Negative Word Cloud:
We were able to find that:
- sentiment analysis not only illustrates how users feel about the app but also highlights actionable areas for improvement
- Instagram feedback is mixed, with more negative sentiment
- 1-star ratings are common, showing frustration despite high engagement
- Users enjoy features such as notes and gif
- country-level shows that Germany reports the lowest satisfaction, in comparison, India and the U.S show the highest
- Despite these complaints, users are active, suggesting Instagram features keep it competive and relevant