Part three: Text Analysis
To understand consumer reviews, the codes below will help understand the review stars distribution and what consumers mentioned frequently about the product.
Step one: Import packages & modules
Step two: data preparation
Step three: review distribution
I separated the text into 3 parts based on the rating stars: Neutral, ratings equal to 3 stars; Positive, ratings greater and equal to 4 stars; Negative, ratings less and equal to 2 stars.
Among 5054 reviews, the average rating score is 4.68/5. The average length of low-rating reviews is 314. The portion of low-rating reviews is 4% which indicates that only a small percentage of customers think the product is performing badly.
Step four: Word Frequency
Let's take an insight into what customers are saying!
I analyzed the positive, negative, and overall ratings by using CountVectorizer to separate the review text into two words and three words. Then I found the top 15 phrases that customers mentioned the most in each rating groups' reviews.
​
Positive ratings:
Most customers think that the Oculus Quest2 VR headset is easy to use and is a VR game suitable for all ages and family gatherings, which is very interesting and brings customers a good VR experience. Some customers also mentioned that the battery life has been longer than expected. A few customers mentioned the need for a Facebook account. Most customers would recommend it to others.
Negative ratings:
Quest2 had a very low percentage of negative reviews, with customers mentioning the need for a Facebook account to sign up for the game the most. A small number of customers also mentioned the need for more clarity.
Overall ratings:
Overall, Customer reviews of Quest 2 have been very good. Most customers found Quest 2 to provide them with a great gaming experience and would recommend it to others.
