Merge pull request #676 from adi271001/ecommerce-trends

Ecommerce trends Analysis

E-Commerce Trends Analysis/Models/README.md

@@ -0,0 +1,104 @@
+# E-commerce Product Sales Prediction
+
+# Models
+
+## Table of Contents
+
+```
+* Goal 
+* Dataset 
+* Description 
+* Libraries 
+* Models and Results 
+* Conclusion
+```
+## Goal
+
+#### To predict sales of e-commerce products using various machine learning models.
+
+## Dataset
+
+#### Link: The dataset is provided within the notebook and contains various e-commerce product details.
+
+## Description
+
+#### * This folder contains the code and resources for predicting sales of e-commerce products using various machine learning models.
+
+#### * The prediction is based on product details such as product name, category, price, rating, number of reviews, stock quantity, discount, and sales.
+
+## Libraries Needed
+
+#### * pandas
+
+#### * numpy
+
+#### * matplotlib
+
+#### * seaborn
+
+#### * plotly
+
+#### * scikit-learn
+
+## Models and Results
+
+#### The project explores the following machine learning models to predict sales:
+
+## 1. Linear Regression
+
+#### Linear Regression is a basic and commonly used predictive analysis model. The model attempts to find the linear relationship between the input features and the target variable (sales).
+
+```
+Results: RMSE: 593.23 R² Score: -0.0170
+```
+### 2. Decision Tree Regressor
+
+#### Decision Tree Regressor builds a model in the form of a tree structure. It breaks down the dataset into smaller subsets while at the same time an associated decision tree is incrementally
+
+#### developed.
+
+```
+Results: RMSE: 855.12 R² Score: -1.1131
+```
+### 3. Random Forest Regressor
+
+#### Random Forest Regressor improves the performance of decision trees by building multiple trees and combining their predictions. It reduces overfitting and improves accuracy.
+
+```
+Results: RMSE: 621.81 R² Score: -0.1173
+```
+### 4. Gradient Boosting Regressor
+
+#### Gradient Boosting Regressor builds an ensemble of trees in a sequential manner, where each tree attempts to correct the errors of the previous one. This model is powerful and effective for regression tasks.
+
+```
+Results: RMSE: 609.70 R² Score: -0.0742
+```
+### 5. Support Vector Regressor (SVR)
+
+#### Support Vector Regressor uses Support Vector Machines for regression tasks. It aims to fit the best line within a threshold value (epsilon) and is effective in high-dimensional spaces.
+
+```
+Results: RMSE: 588.42 R² Score: -0.0006
+```
+
+### 6. Logistic Regression
+
+#### Logistic Regression is typically used for classification tasks, but here it was included for comparative purposes. Its performance indicates it is not suitable for regression tasks like sales prediction.
+
+```
+Results: Accuracy: 0.495
+```
+
+* ![Accuracy Comparison](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___42_0.png?raw=true)   
+* ![Logistic REgression Accuracy](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___43_0.png?raw=true)   
+
+
+## Conclusion
+
+#### * Among the models tested, the Support Vector Regressor (SVR) performed the best with the lowest RMSE (588.42) and the least negative R² score (-0.001), making it the most accurate model
+for predicting e-commerce product sales in this analysis.
+
+#### * Best Performing Model: Support Vector Regressor (SVR)
+
+#### * Next Steps: Consider tuning hyperparameters and exploring other algorithms for potentially better performance.

E-Commerce Trends Analysis/README.md

@@ -0,0 +1,168 @@
+# E-commerce Product Trends Analysis
+
+## Table of Contents
+
+```
+* Goal 
+* Dataset
+* Description 
+* What I Had Done 
+* Installation 
+* Libraries 
+* EDA Results
+* Models and Results 
+* Conclusion 
+* Contributing 
+* Signature
+```
+## Goal
+
+#### To analyze e-commerce product trends and predict sales using various machine learning models.
+
+## Dataset
+
+#### Link: https://www.kaggle.com/datasets/muhammadroshaanriaz/e-commerce-trends-a-guide-to-leveraging-dataset
+
+#### details.
+
+## Description
+
+#### * This folder contains the code and resources for analyzing e-commerce product trends and predicting sales using various machine learning models.
+
+#### * The analysis is based on product details such as product name, category, price, rating, number of reviews, stock quantity, discount, and sales.
+
+## What I Had Done
+
+## Installation
+
+#### Clone the repository using the following command:
+
+```
+git clone https://github.com/yourusername/ecommerce-product-trends.git cd
+ecommerce-product-trends
+```
+#### To run the notebook and reproduce the results, you need to have Python installed along with the necessary libraries. You can install the required libraries using the following
+
+#### command:
+
+```
+pip install -r requirements.txt
+```
+#### Run the Jupyter notebook:
+
+```
+jupyter notebook ecommerce-trends-eda-models.ipynb
+```
+## Libraries Needed
+
+#### * pandas
+
+#### * numpy
+
+#### * matplotlib
+
+#### * seaborn
+
+#### * plotly
+
+#### * scikit-learn
+
+
+## Exploratory Data Analysis Results
+
+#### * The dataset contains a wide range of product categories with varying prices, ratings, number of reviews, stock quantities, discounts, and sales.
+
+#### * Initial visualizations indicate significant trends and correlations among these features.
+
+### Graphs and Analysis
+
+#### 1. Relationship Graphs
+
+#### Insights: There are clear trends between price and sales, rating and sales, and discount and sales.
+
+#### 2. Cluster Graph
+
+#### Insights: Products are clustered into different groups based on their features, which helps in segmenting the data.
+
+#### 3. Correlation Matrix
+
+#### Pearson correlation Matrix
+
+#### Insights: The Pearson correlation matrix shows the linear correlation between different features.
+
+#### 4. Predictive Power Score
+
+#### Insights: This score helps in identifying the predictive power of different features for the target variable.
+
+#### 5. Line of Best Fit Graphs
+
+#### Insights: These graphs show the trends and best fit lines for key relationships in the data.
+
+* ![description of dataset](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___7_1.png?raw=true)
+* ![Distribution of ratings](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___8_1.png?raw=true)
+* ![Distribution of other features](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___9_1.png?raw=true)
+* ![correlation matrix](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___13_0.png?raw=true)
+* ![Top 10 Products](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___15_0.png?raw=true)
+* ![price vs sales clustering graph](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___17_0.png?raw=true)
+* ![pairplot of numerical features](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___18_2.png?raw=true)
+* ![Word Cloud](https://github.com/adi271001/ML-Crate/blob/ecommerce-trends/E-Commerce%20Trends%20Analysis/Images/__results___31_1.png?raw=true)   
+
+## Models and Results
+
+#### The project explores the following machine learning models to predict sales:
+
+### 1. Linear Regression
+
+```
+Results: RMSE: 593.23 R² Score: -0.0170
+```
+### 2. Decision Tree Regressor
+
+```
+Results: RMSE: 855.12 R² Score: -1.1131
+```
+### 3. Random Forest Regressor
+
+```
+Results: RMSE: 621.81 R² Score: -0.1173
+```
+### 4. Gradient Boosting Regressor
+
+```
+Results: RMSE: 609.70 R² Score: -0.0742
+```
+### 5. Support Vector Regressor (SVR)
+
+```
+Results: RMSE: 588.42 R² Score: -0.0006
+```
+### 6. Logistic Regression
+
+```
+Results: Accuracy: 0.495
+```
+## Conclusion
+
+#### * Based on the evaluation of various machine learning models, the Support Vector Regressor (SVR) emerged as the best-performing model with the lowest RMSE (588.42) and the least negative R² score (-0.001), indicating it provides the most accurate predictions among the tested models.
+
+#### * Best Performing Model: Support Vector Regressor (SVR) due to its lowest RMSE and R² scores.
+
+#### * Next Steps: Verify the data processing steps and re-evaluate the models to identify any issues. Consider tuning model hyperparameters and exploring other algorithms for improved performance.
+
+#### * Important Features: Features such as price, rating, and discount were expected to be influential in predicting sales.
+
+## Contributing
+
+#### Contributions are welcome! Please read the contribution guidelines first.
+
+## Signature
+
+#### Aditya D
+
+#### Github: https://www.github.com/adi271001
+
+#### LinkedIn: https://www.linkedin.com/in/aditya-d-23453a179/
+
+#### Topmate: https://topmate.io/aditya_d/
+
+#### Twitter: https://x.com/ADITYAD29257528

E-Commerce Trends Analysis/Results/model_results.csv

@@ -0,0 +1,7 @@
+Model,RMSE,R^2 Score,Accuracy
+Linear Regression,593.2257019066986,-0.01700355866153469,
+Decision Tree,855.1219971442671,-1.1131908015070708,
+Random Forest,621.8057745031804,-0.11735726372963051,
+Gradient Boosting,609.695323177347,-0.07425722239816679,
+SVR,588.4236152648417,-0.0006051707827330333,
+Logistic Regression,,,0.495

E-Commerce Trends Analysis/requirements.txt

@@ -0,0 +1,7 @@
+pandas==1.3.3
+numpy==1.21.2
+matplotlib==3.4.3
+seaborn==0.11.2
+plotly==5.3.1
+scikit-learn==0.24.2
+