SC1015 Customer Personality Analysis

About

This is a Mini-Project for the course SC1015 Introduction to Data Science and Artificial Intelligence conducted by Nanyang Technological University's College of Computing and Data Science.

The dataset used in this project is extracted from Customer Personality Analysis on Kaggle.

The source codes, in order, are as follows:

1. Data Exploration
2. Data Cleaning
3. Data Visualization
4. Data Splitting
5. Linear Regression
6. Poisson Regression

Contributors

• Lim Boon Hian - Data Exploration, Data Cleaning, Data Visualization
• Lim Dong Wan - Linear Regression, Data Splitting
• Marvell Ung Wew - Poisson Regression, Grid Search Cross Validation

Problem Definition

Predicting customer sales using customer information

Variables in Focus

Response variables: TotalPurchase, MntGroceryProducts, MntWines, MntGoldProds

Predictor variables:

• Categorical:
  • Education
  • Marital_Status
  • HaveChild
  • YearRange
• Numerical:
  • Income
  • TotalChild
  • NumWebVisitsMonth

Machine Learning Models Used

1. Linear Regression
2. Poisson Regression

Insights and Recommendations

Main insight: The company utilizes customer-focused sales tactics with customer’s income as a guideline.

Recommendations:

1. Target audience should be customers who have / are:
   • fewer children
   • higher income
   • higher levels of education
   • born in 1940 - 1950
2. Improve online sales by upgrading company website, so as to make it more attractive and appealing to customers.

Conclusion

1. Customer income serves as the most important predictor to predict customer expenditure
2. Linear Regression yield better results than Poisson Regression
3. GridSearch CV only marginally improved the Linear Regression model

Things Learnt

1. K-fold splitting
2. Poisson Regression
3. Plotly subplot
4. Grid Search Cross Validation
5. Altair interactive plot
6. Get dummy values for categorical variables in Linear Regression
7. Collaborating on Github :)

References

• https://www.kaggle.com/datasets/imakash3011/customer-personality-analysis (Original Dataset)
• https://ntulearn.ntu.edu.sg/webapps/blackboard/content/listContent.jsp?course_id=_2606895_1&content_id=_2762960_1
• https://stackoverflow.com/questions/21912634/how-can-i-sort-a-boxplot-in-pandas-by-the-median-values
• https://altair-viz.github.io/gallery/index.html
• https://www.kaggle.com/code/jnikhilsai/cross-validation-with-linear-regression/notebook
• https://towardsdatascience.com/gridsearchcv-for-beginners-db48a90114ee
• https://machinelearningmastery.com/rfe-feature-selection-in-python
• https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
• https://medium.com/lcc-unison/how-to-poisson-regression-model-python-implementation-1c672582eb96
• https://timeseriesreasoning.com/contents/poisson-regression-model
• https://matplotlib.org/stable/gallery/subplots_axes_and_figures/subplots_demo.html
• https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.KFold.html
• https://medium.com/analytics-vidhya/implementing-linear-regression-using-sklearn-76264a3c073c
• https://machinelearningmastery.com/rfe-feature-selection-in-python/