FRESCO-RETAIL-PROJECT

The data analyses Fresco Retail’s customer's transaction data to predict return decision using various data points like customer background, payment modes, store types, product nature, etc.

Objective

• To study the influence of different transaction data points and their impact on a customer’s final decision to return the products bought in a transaction.

Variable Description:

• transaction_id: Unique transaction number for each sale
• Month_code: The month of transaction
• prod_cat_code: Product category code. A product may have multiple sub-categories
• prod_cat: Product category name mapped to prod_cat_code.
• prod_subcat_code: Product sub-category code
• prod_subcat: Product sub-category name mapped to product_subcat_code
• Qty: Quantity of products bought / returned
• Rate: Price per unit of a product in local currency
• Amount: Total amount (Qty * Rate) without the taxes
• Tax: Tax amount by local government
• Delivery_chrgs: Delivery charges to deliver the products ordered
• Payment_mode: Mode of payment used by the customer to pay for a transaction
• Store_type: Type of store where the transaction took place. eShop means online commerce. MBR means Multi branded retail store. Flagship store means standalone fully owned shop by the retailer. Teleshop means orders received on retailer phone numbers.
• Reviews: Review left by the customer on the product or transaction
• Cust_id: Customer identification number
• DOB: Customer’s date of birth
• Gender: Customer’s gender
• Education_Level_Code: Code assigned to different education levels of a customer
• Level_Education: Education level mapped to Education_Level_Code
• Profession_Code: Customer’s profession code
• Profession Type: Type of profession (Salaried, Self Employed, Others)
• City_code: City identification code assigned to each city
• RETURN: This is our target variable that stores the response of whether the transaction was finally returned by the customer

Approach & Steps:

1. Understand the data variables properly. Refer to above the variable description.
2. Clean the data: Clean the data, that is, fill the missing values (if any), treat the outliers (or odd values), etc. Ensure each variable’s data is as per the nature of the variable (eg – Date field should contain only date values, numeric column should be formatted as numeric, etc.).
3. Conduct EDA (Exploratory Data Analysis) on the cleaned Data: Summarize, explore the data and then decide your strategy. Make note of any important assumptions that you make.
4. Uni-variate and Bi-variate Analysis: Check the distribution of independent variables and also compare them with the dependent variable.
5. Feature Engineering: Create new meaningful features based on the existing features by applying some aggregation functions on them.
6. Identify the most important variables (or data parameters) that affect the final decision: Identify the impact of each variable on the final result graphically (correlation / scatter plots, regression plots, etc.). Keep those variables that affect the final outcome.
7. Develop and Validate Samples: Divide samples into 2 parts: Development Sample (70%) & Validation Sample (30%). Build your analysis model using the Development Sample, and validate it on the validation sample and then predict on test sample.
8. Model Building: Analyze the dependent variable and decide which technique out of regression or classification to use and hence build the model.
9. Improving model accuracy: Perform various iterations by eliminating or adding the variables to see if the model accuracy is improving or not. Also, you can apply various transformation like log transformation on dependent variable or independent variables or both to improve accuracy.
10. Model Comparison: Comparing the chosen model with other similar models that could have been used in this project

Dependencies

• numpy
• matplotlib
• scikit-learn
• seaborn
• pyhthon 3.9

Usage

Just run jupyter notebook in terminal and it will run in your browser.

Install Jupyter here i've you haven't.

Models with their Accuracy of Prediction

• Logistic Regression Model: The Model accuracy is 0.903

             precision   recall   f1-score   support
       0       0.98      0.91      0.94      5236
       1       0.53      0.88      0.66       616
  

• Random Forest Model: The Model accuracy is 0.94

             precision  recall   f1-score   support
      0       0.97      0.97      0.97      5236
      1       0.75      0.72      0.73       616
  

• If we compare logistic model with random forest the accurcay and precision is quite much better than logistic model. It mean random forest model is better.

• https://github.com/rahkum96/FRESCO-RETAIL-PROJECT/blob/main/FRESCO%20RETAIL.ipynb

Steps to run this model in your system

• Clone or download the repo.
• Open command prompt in the downloaded folder.

PowerPoint presenation

Downlaod the ppt from this link https://github.com/rahkum96/FRESCO-RETAIL-PROJECT/blob/main/FRESCO%20RETAIL-%20PROJECT.pptx

Dataset:

• Data set is already uploaded
• https://github.com/rahkum96/FRESCO-RETAIL-PROJECT/blob/main/Fresco_LargeRetailer.xlsx