Aerofit Treadmill Customer Profiling

Aerofit Treadmill
Customer Profiling

Exploratory data analysis to build customer profiles for each treadmill model — helping Aerofit's sales team recommend the right product to the right buyer.

TypeEDA & Probability Analysis

DomainFitness / Retail

Dataset180 customers · 9 features

ToolsPython · Pandas · Seaborn

CourseScaler Academic Case Study

Who buys which treadmill?

Aerofit sells three treadmill models — KP281 (entry-level), KP481 (mid-range), and KP781 (premium). The fitness retail company wants to understand who buys what — so their sales team can make data-driven recommendations during purchase consultations.

🎯

The goal

Build customer profiles for each treadmill model using age, income, fitness level, gender, and usage frequency. Give the sales team a clear picture of the typical buyer for each product.

What we're working with

Feature	Type	Description
Product	Categorical	KP281 / KP481 / KP781
Age	Numerical	Customer age in years
Gender	Categorical	Male / Female
Education	Numerical	Years of education
MaritalStatus	Categorical	Single / Partnered
Usage	Numerical	Planned weekly treadmill uses
Fitness	Numerical (1-5)	Self-rated fitness level
Income	Numerical	Annual income in USD
Miles	Numerical	Expected weekly miles

How I built the profiles

Univariate Analysis

Distributions for age, income, fitness, usage, miles. Identified that income and miles are right-skewed with outliers.

histplot · boxplot

Bivariate Analysis by Product

Cross-tabulated each feature against product type. Used countplots, boxplots, and mean comparisons to isolate model-specific patterns.

groupby · crosstab · boxplot(hue)

Correlation Analysis

Built correlation heatmap to understand feature relationships. Income–Education and Usage–Miles are the strongest correlations.

heatmap · pairplot

Marginal & Conditional Probability

Calculated P(Product) and P(Feature | Product) to build probabilistic customer profiles for each treadmill.

value_counts(normalize=True)

Python — customer_profiling.py

# Conditional probability: P(Gender | Product)
pd.crosstab(df['Product'], df['Gender'], normalize='index')

# Mean stats per product
df.groupby('Product')[['Age','Income','Fitness','Usage','Miles']].mean()

# Output:
#          Age    Income  Fitness  Usage  Miles
# KP281   28.5   46,400    3.0    3.3    82
# KP481   28.9   48,900    3.1    3.5    87
# KP781   29.1   58,500    4.2    4.8   166

The three buyer types

KP281

Entry Level · Most Popular
Age ~28 · Income ~$46K
Fitness: 3/5 · Miles: 82/wk

KP481

Mid Range · Balanced
Age ~29 · Income ~$49K
Fitness: 3.1/5 · Miles: 87/wk

KP781

Premium · Power Users
Age ~29 · Income ~$58K
Fitness: 4.2/5 · Miles: 166/wk

💰

Income is the clearest separator

KP781 buyers earn ~25% more than KP281 buyers. Income is the strongest predictor of product tier.

🏃

Miles run per week is the key usage signal

KP781 users plan to run 166 miles/week vs 82 for KP281. Usage intensity is a critical profiling dimension.

⚡

Fitness level drives premium purchase

KP781 buyers rate themselves 4.2/5 vs 3.0/5 for KP281. Premium treadmills are bought by already-fit customers.

👫

Gender and marital status matter less

Males slightly prefer KP781. Partnered individuals lean toward mid-range. But these are weak signals compared to income and fitness.

What Aerofit should do with this

Build a recommendation system

Ask 3 questions at purchase: fitness level (1–5), weekly budget for exercise, and income range. These three alone predict the right product.

Upsell KP281 buyers aggressively

KP281 buyers are younger and lower-income — but if they become regular users, they're prime upgrade candidates in 12–18 months.

Target KP781 via fitness communities

Premium treadmill buyers already have high fitness ratings. Gym partnerships, running apps, and fitness influencer channels are the right media channels.

Aerofit Treadmill
Customer Profiling

Who buys which treadmill?

What we're working with

How I built the profiles

The three buyer types

What Aerofit should do with this