Predicting BMW Sales Classification (2010–2024) Using Logistic Regression

Overview

To understand which vehicle attributes most influence whether BMW sales fall into high or low performance categories, a Logistic Regression classifier was trained on the dataset’s numerical and categorical variables using a Pipeline structure.

The model automatically scaled numeric features and encoded categorical ones, ensuring balanced and unbiased learning across all predictors.

Model Framework

Pipeline Components

1. Preprocessing (ColumnTransformer)

   • Numerical features: Scaled using StandardScaler()

   • Categorical features: Encoded using OneHotEncoder()

2. Estimator: LogisticRegression() (solver: lbfgs, max_iter=1000)

3. Cross-Validation Accuracy: 0.9938 ± 0.00067

Model Performance

Confusion Matrix

$$\begin{bmatrix} 3049 & 0 \\ 58 & 6893 \end{bmatrix}$$

• Only 58 misclassifications out of 10,000 test records.

• No false positives for the “High” class, indicating excellent precision and generalization.

Interpretation

1. Model Strength:
   The logistic regression model achieved near-perfect performance, confirming strong predictive relationships between key features (e.g., Price, Engine_Size, Sales_Volume, Mileage) and sales category.

2. High Recall:
   The model identifies nearly all high-sales cases, minimizing missed opportunities in predicting premium market segments.

3. Feature Insight:

   • Sales_Volume and Price_USD were dominant predictors.

   • Region and Fuel_Type added contextual strength, aligning with global sales diversity.

4. Strategic Use:

   • BMW or dealers can use such models to forecast sales potential for new models.

   • Governments or industry analysts can assess luxury car demand elasticity and policy impacts on premium automotive markets.

Policy and Industry Implications

• Market Forecasting: Predictive analytics helps manufacturers align production with demand surges or slumps.

• Sustainability Goals: Including Fuel_Type and Engine_Size in the prediction enhances understanding of green technology adoption patterns.

• Revenue Planning: Accurate classification supports pricing and dealership inventory management.

Acknowledgments

• Dataset: BMW Sales Data (2010–2024), analyzed within the DatalytIQs Academy Analytics Framework.

• Tools & Libraries: Python (scikit-learn, pandas, numpy, matplotlib).

• Contributors:

  • Collins Odhiambo Owino — Lead Analyst & Author, DatalytIQs Academy

  • Kaggle Automotive Datasets — Data structuring and reference

  • BMW Group Annual Reports (2010–2024) — Validation context for market trends

Author’s Note

Written by Collins Odhiambo Owino
Founder & Lead Researcher, DatalytIQs Academy
Empowering learners and professionals in Mathematics, Economics, and Finance through data-driven insights.