Z-ENG Data-Driven E-Commerce Analytics: ETL, Business Insights, and Predictive Modeling
2025-2026 tavasz
Nincs megadva
Téma leírása
As the first step in the thesis project, the focus will be on performing Extract, Transform, Load (ETL) operations and thorough data preprocessing. This includes handling missing values, detecting and treating outliers, and resolving inconsistencies in the dataset. A key task will be converting combined order date and time fields into a standardized datetime format. Additionally, new derived columns will be created—for example, calculating Revenue by subtracting discounts from Sales. After data cleaning and transformation, comprehensive Exploratory Data Analysis will be conducted to uncover actionable business insights.
This phase will include:
-Trend Analysis: Identifying weekly and monthly sales trends to understand seasonality and growth.
-Customer Behavior Analysis: Examining purchasing patterns based on gender and device type to reveal consumer preferences.
-Profitability Analysis: Determining the best-selling products and most profitable categories to guide business strategy.
To effectively communicate findings, advanced visualizations will be developed using tools such as Power BI, Tableau, or Python Dash. These will include:
-Heatmaps to visualize peak shopping hours and highlight customer engagement windows. Time-series charts to display sales trends over time.
-Interactive dashboards for customer segmentation and profitability analysis.
To enrich the project, data science techniques will be integrated into the analysis:
-Trend and Seasonality Detection: Time-series analysis will be used to identify recurring seasonal patterns in sales data.
-Customer Segmentation: Using K-Means clustering, customers will be grouped based on features like total sales, quantity purchased, device type, and profitability. This will help distinguish high-value customers from low-value ones.
-Sales/Profit Prediction: Predictive models such as Linear Regression and RandomForest will be implemented to forecast future sales or profits.
-Feature Engineering: Key features will be derived to improve model accuracy, including the impact of discounts, shipping costs, and seasonal variations.
Tasks to be performed by the student will include:
-ETL, Data Cleaning, and Preprocessing
-Exploratory Data Analysis (EDA) and Business Insights
-Advanced Visualizations
-Adding Data Science Components
Maximális létszám:
1 fő