HeyMrDigital
E-commerce Game Recommendation System

E-commerce Game Recommendation System Python Project

A Content-Based Recommender Using NLP and Cosine Similarity
← Back to Portfolio

Project Summary

The Product Recommendation System is a content-based recommender engine that suggests similar products based on their textual features such as category, subcategory, and type. Designed to enhance user engagement and improve discovery in e-commerce or inventory platforms, it uses natural language processing (NLP) and vector-based similarity techniques to generate recommendations.

Key Features & Responsibilities

  • 📚 Data Preprocessing & Cleaning: Loaded and cleaned a real-world product dataset, standardized text, engineered new features.
  • 📊 Exploratory Data Analysis: Visualized category/type distributions, identified top product types.
  • 🔍 Text Vectorization: Applied TF-IDF and CountVectorizer to represent product features as vectors.
  • 🤖 Recommendation Engine: Built two recommenders (TF-IDF and soup-based), used cosine similarity for relevance.
  • 🔄 Evaluation & Comparison: Compared outputs, demonstrated improvement with soup-based model.

Code Showcase

Data Cleaning & Feature Engineering
df = pd.read_csv('productlist.csv')
df = df.dropna()
for col in ['category', 'sub_category', 'type']:
    df[col] = df[col].apply(lambda a: list(map(lambda x: x.strip(), re.split('& |, |*|
', a))))
def cleaner(x):
    if isinstance(x, list):
        return [str.lower(i.replace(' ', '')) for i in x]
    elif isinstance(x, str):
        return str.lower(x.replace(' ', ''))
    else:
        return ''
for col in ['category', 'sub_category', 'type']:
    df[col] = df[col].apply(cleaner)
def couple(x):
    return ' '.join(x['category']) + ' ' + ' '.join(x['sub_category']) + ' ' + ' '.join(x['type'])
df['soup'] = df.apply(couple, axis=1)
TF-IDF & Cosine Similarity
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import linear_kernel

tfidf = TfidfVectorizer(stop_words='english')
tfidf_matrix = tfidf.fit_transform(df['product'])
cosine_sim = linear_kernel(tfidf_matrix, tfidf_matrix)
Recommendation Function
indices = pd.Series(df.index, index=df['product']).drop_duplicates()
def get_recommendations(title, cosine_sim=cosine_sim):
    idx = indices[title]
    sim_scores = list(enumerate(cosine_sim[idx]))
    sim_scores = sorted(sim_scores, key=lambda x: x[1], reverse=True)
    sim_scores = sim_scores[1:11]
    product_indices = [i[0] for i in sim_scores]
    return df['product'].iloc[product_indices]

Results & Value Delivered

  • Improved product discovery and user engagement for e-commerce platforms
  • Enabled scalable, content-based recommendations using NLP
  • Delivered clean, modular code and reusable recommendation functions
  • Provided actionable insights through EDA and visualizations