Omarrran/50k_IMBD_Movie_Review_by_HNM

🎬 50K IMDB Movie Reviews Dataset

A balanced sentiment analysis dataset with 50,000 IMDB movie reviews

Overview • Dataset Structure • Statistics • Usage • Citation

---

📖 Overview

This dataset contains 50,000 movie reviews from IMDB, perfectly balanced between positive and negative sentiments. Each review includes the original text, reviewer rating (1-10), sentiment label, and source URL, making it ideal for sentiment analysis, text classification, and NLP research.

✨ Key Features

Feature	Description
🎯 Perfectly Balanced	25,000 positive + 25,000 negative reviews
⭐ Rating Included	Original 1-10 reviewer ratings preserved
🔗 Source URLs	Links to 7,036 unique movies
📝 Rich Text	Average 227 words per review
✅ Clean Data	Zero missing values

---

📊 Dataset Structure

Data Fields

Column	Type	Description
review	string	Full text of the movie review
movie_url	string	IMDB URL of the reviewed movie
reviewer_rating	int64	Rating given by reviewer (1-10 scale)
label	string	Sentiment label (Positive or Negative)

Sample Data

{
    "review": "This movie was absolutely fantastic! The acting was superb...",
    "movie_url": "https://www.imdb.com/title/tt0111161/",
    "reviewer_rating": 9,
    "label": "Positive"
}

---

📈 Statistics

Basic Information

Metric	Value
Total Reviews	50,000
Total Columns	4
Memory Usage	74.68 MB
Missing Values	0 (Clean!)

Label Distribution

┌─────────────────────────────────────────────────────────┐
│  Negative  ████████████████████████████████████  25,000 (50%)
│  Positive  ████████████████████████████████████  25,000 (50%)
└─────────────────────────────────────────────────────────┘

Label	Count	Percentage
Negative	25,000	50.0%
Positive	25,000	50.0%

⭐ Reviewer Rating Statistics

Statistic	Value
Mean	5.50
Std Dev	3.48
Min	1
Max	10
Median	5.5

Rating Distribution

Rating	Count	Visualization
1	10,122	████████████████████
2	4,586	█████████
3	4,961	██████████
4	5,331	███████████
7	4,803	██████████
8	5,859	████████████
9	4,607	█████████
10	9,731	███████████████████

📝 Review Text Statistics

Metric	Characters	Words
Average	1,285	227
Minimum	32	4
Maximum	13,584	2,450

🎬 Movie Coverage

Metric	Value
Unique Movies	7,036
Avg Reviews per Movie	7.11

---

🚀 Usage

Loading with Hugging Face Datasets

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("Omarrran/50k_IMBD_Movie_Review_by_HNM")

# Access the data
train_data = dataset['train']

# View first example
print(train_data[0])

Loading with Pandas

import pandas as pd
from datasets import load_dataset

dataset = load_dataset("Omarrran/50k_IMBD_Movie_Review_by_HNM")
df = dataset['train'].to_pandas()

# Explore
print(f"Shape: {df.shape}")
print(df.head())

Quick Sentiment Classification Example

from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report

# Prepare data
X = df['review']
y = df['label']

# Split
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42, stratify=y
)

# Vectorize
vectorizer = TfidfVectorizer(max_features=10000, stop_words='english')
X_train_vec = vectorizer.fit_transform(X_train)
X_test_vec = vectorizer.transform(X_test)

# Train
clf = LogisticRegression(max_iter=1000)
clf.fit(X_train_vec, y_train)

# Evaluate
print(classification_report(y_test, clf.predict(X_test_vec)))

Fine-tuning Transformers

from transformers import AutoTokenizer, AutoModelForSequenceClassification
from transformers import TrainingArguments, Trainer

# Load model and tokenizer
model_name = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=2)

# Tokenize dataset
def tokenize_function(examples):
    return tokenizer(examples["review"], padding="max_length", truncation=True)

tokenized_dataset = dataset.map(tokenize_function, batched=True)

# Train (simplified)
training_args = TrainingArguments(
    output_dir="./results",
    num_train_epochs=3,
    per_device_train_batch_size=16,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_dataset["train"],
)

trainer.train()

---

🎯 Suitable Tasks

Task	Description
🔍 Sentiment Analysis	Binary classification (Positive/Negative)
📊 Rating Prediction	Predict 1-10 rating from review text
📝 Text Classification	Multi-class or multi-label classification
🤖 Language Modeling	Fine-tune GPT/BERT on movie domain
📈 Feature Extraction	Extract embeddings for downstream tasks
🔬 NLP Research	Benchmark models on balanced dataset

---

⚠️ Limitations

• English language only
• Movie reviews domain (may not generalize to other domains)
• User-generated content with varying quality
• Temporal bias based on collection period
• Ratings 5 and 6 are underrepresented (polarized dataset)

---

📜 Citation

If you use this dataset in your research, please cite:

@dataset{50k_IMDB_Movie_Review_by_HNM,
  title     = {50K IMDB Movie Reviews Dataset},
  author    = {Haq Nawaz Malik},
  year      = {2025},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/Omarrran/50k_IMBD_Movie_Review_by_HNM}
}

---

📄 License

This dataset is provided for research and educational purposes under the MIT License.

---

Created by Haq Nawaz Malik

⭐ If you find this dataset useful, please consider giving it a star!