Situation
This is the continuation of Luxura Analysis Part 1. In this project, we continue our analysis on an individual customer’s purchase data to provide personalized insights.
Backround
Luxura is an e-commerce platform selling high-end fashion brands. Mr. Chu is one of their top customers and investors who heavily purchases from his favorite brand on Luxura, Adibi.
However, Mr. Chu wants to better manage his expenses going forward. He has spent a significant portion of his luxury budget on Adibi items. By analyzing his detailed historical purchases from Luxura and building a personalized predictive model, Mr. Chu can receive tailored insights on his buying behavior and data-backed forecasts of future Adibi order amounts. These custom analytics will empower Mr. Chu to more strategically plan his fashion budget and purchase levels.
Providing individualized recommendations and predictions based on customer data is also strategically important for Luxura. These personalized services increase engagement, satisfaction, and retention with high-value users like Mr. Chu.
Problem Definition
The problem was providing tailored analytics on Mr. Chu’s purchasing behavior and predicting his future Adibi orders to enable better expense planning.
Objective
The objective was to analyze Mr. Chu’s personal order data to identify key drivers of and predict his Adibi purchases to provide custom insights.
Task & Action
| Task | Action | Reason |
|---|---|---|
| Correlation analysis | Calculated correlations between variables and Adibi order value | Identified variables strongly related to target metric |
| Construct regression model | Built regression model to predict customer’s Adibi order value based on characteristics | Regression model used as main objective to construct the regression analysis |
| Regression modeling | Performed backward stepwise regression to determine factors with biggest effect on Adibi order value | Found key drivers of high Adibi spending |
| Present insights | Delivered personalized analytics based on intperpretation of regresion result | Provided custom insights for improved expense planning |
| Make recommendations | Recommend Mr.Chu what he should do to change his behavior on shopping from indicators already determined from regression result | Develop recommendations to help Mr.Chu future behavior |
Result - Slide Deck
By completing this project’s, I’ve gained crucial hands-on experience to provide insight for client by apply correlation analysis and regression analysis. I’ve enhanced my skills from those activity, that very crucial in the field of data analytics.
In summary, I’ve created 2 models and apply backward stepwise regression for all models. I found that Mr. Chu can utilize Celinna value order can be predictor to Mr. Chu Adibi Spending Behavior. Due to the poor quality of the data, I have minimised the violation of regression assumptions, so the model is feasible to use as a form of recommendation. You can see more detailed result:
Attachments
» Dataset Link « | » Spreadsheet Link « | » Google Colab Link «
I’m grateful you took the time to explore my portfolio site and view my work. Your interest means a lot.
Feel free to provide any feedback or ask questions - I welcome your thoughts. Thank you again for checking out my projects and considering my skills. I appreciate you taking the time to do so.