Tata Motors · Customer Analytics

Hybrid Vehicle Purchase Prediction

AI-powered customer purchase prediction for Tata Motors

Identify customers who are more likely to purchase a hybrid vehicle using demographic, income and vehicle ownership data, scored by a logistic regression model trained on Tata Motors' historical customer records.

Enter Customer Details step 1
Actual age range — mapped internally to the model's age band.
Highest qualification — mapped internally to the model's education level.
Enter income directly in Lacs, as recorded in the training data.
Number of vehicles currently owned by the household (0–4).
⚠ Please select all four fields before predicting.
Prediction Result step 2
No prediction yet
Fill in the customer details and click Predict Purchase to run the trained model.

Purchase Prediction

—

—
purchase probability
Logistic function σ(z)z = —

04 / Methodology

How the Prediction Works

01

Customer Data

Age, Education, Income and Vehicle Ownership are collected using the same encoding as the original training dataset.

02

Logistic Regression Model

A binary classifier — trained with scikit-learn's LogisticRegression — computes a weighted score and converts it to a purchase probability.

03

Purchase Prediction

The customer is classified as a Potential Buyer or Not Immediate Buyer based on the model's decision threshold.

The model analyzes Age, Education, Income and Vehicle Ownership to classify customers into two groups: Potential Buyer or Not Immediate Buyer. Logistic Regression is used here specifically as a binary classification model — it does not output a raw quantity, it outputs the probability of belonging to the "purchase" class.

INPUT VARIABLES
  • Age
  • Education
  • Income
  • Vehicle Ownership
OUTPUT
  • Potential Buyer
  • Not Immediate Buyer

05 / Application

Business Use Cases

Targeted Marketing

Identify potential customers for targeted hybrid vehicle campaigns.

Customer Segmentation

Segment customers according to predicted purchase intent.

Sales Prioritization

Help sales teams prioritize customers classified as potential buyers.

Campaign Optimization

Use prediction results to improve marketing outreach.

06 / Model Information

Trained Model Details

Algorithm
Logistic Regression
Features
4
Target
Purchase hybrid vehicle
Classes
Buyer / Not immediate
Training Accuracy
89.26%
Training Records
475

Coefficients (Age: 0.7986, Education: 0.4508, Income: 0.7236, Vehicle Ownership: 1.0570, Intercept: −11.1288) and the 89.26% training accuracy shown above were computed by fitting scikit-learn's LogisticRegression on the actual uploaded training dataset (475 records) — no metric on this page is fabricated. Age and Education are collected as real-world values (age range, qualification) and converted to the dataset's coded bands (Age 1–5, Education 1–4) before scoring; Income is entered directly in Lacs, matching the scale the model was trained on. This demo evaluates the fitted equation directly in the browser; the same coefficients can be served from a persisted model file (joblib/pickle) behind a POST /predict FastAPI endpoint in production, as outlined in the README.