How I Extracted Amazon Review Text Data from Kaggle's .bz2 Dataset for Sentiment Analysis

I recently started working on my first NLP and Sentiment Analysis project using the Amazon Reviews dataset on Kaggle.

I expected to find a normal CSV file that I could load using pandas.read_csv().

Instead, I found files like:

train.ft.txt.bz2

test.ft.txt.bz2

At first, I was confused.

What is a .bz2 file?

How do I read it?

Why isn't there a CSV file?

After some searching and experimentation, I discovered that the dataset is stored in a compressed BZip2 format. Fortunately, Python provides a built-in bz2 library that can read these files directly.

Step 1: Reading the compressed file

import bz2

import pandas as pd

file_path = "/kaggle/input/datasets/bittlingmayer/amazonreviews/train.ft.txt.bz2"

with bz2.open(file_path, "rt", encoding="utf-8") as f:

lines = f.readlines()

print(lines[0])

The output looked something like this:

label2 Stunning even for the non-gamer: This soundtrack was beautiful...

At this point, I noticed that every line starts with a label followed by the review text.

Step 2: Separating labels and reviews

labels = []

sentences = []

for line in lines:

label = 1 if line.split(" ")[0] == "label2" else 0

sentence = line.split(" ", 1)[1]

labels.append(label)

sentences.append(sentence)

The dataset uses:

label1 → Negative review

label2 → Positive review

Since machine learning models work better with numbers, I converted them into:

0 → Negative

1 → Positive

Step 3: Creating a DataFrame

df = pd.DataFrame({

"review": sentences,

"sentiment": labels

})

print(df.head())

Output:

                                          review  sentiment

0 Stunning even for the non-gamer: This soundtra... 1

1 The best soundtrack ever to anything... 1

2 Amazing! This soundtrack is my favorite music... 1

Now the dataset is finally in a format that can be used for preprocessing, visualization, and model training.

What I Learned

Not every Kaggle dataset comes as a CSV file.

.bz2 is simply a compressed file format.

Python's built-in bz2 library can read these files directly.

Amazon Review datasets use text labels instead of numeric labels.

Converting the data into a DataFrame makes the next NLP steps much easier.

Final Thoughts

This was a small issue, but it completely blocked my progress for a while.

As a beginner in NLP, I am discovering that many challenges are not about machine learning algorithms but about understanding datasets and data formats.

Hopefully this saves someone else a few hours of confusion.