📊 UserInsights — Web Analytics & User Behavior Analysis

End-to-end customer analytics pipeline — from synthetic data generation to interactive dashboards.

Analyze user behavior, cohort retention, conversion funnels, and RFM customer segments with real-time filters.

Features • Quick Start • Dashboard • Project Structure • Key Findings • Documentation

---

✨ Features

Feature	Description
Synthetic Data Generator	Generates realistic e-commerce event data (34K+ events, 10K users, 6 months) with configurable conversion probabilities
Exploratory Data Analysis	7 automated plots covering funnel, daily trends, hourly heatmap, channel performance, device breakdown, and more
Cohort Analysis	Monthly retention cohorts with interactive heatmaps, revenue-per-user analysis, and retention curves
Conversion Funnel	5-step funnel (visit → view → cart → checkout → purchase) with sankey diagrams and channel-level breakdowns
RFM Segmentation	Recency-Frequency-Monetary scoring with 10 customer segments — identifies Champions, Loyal Customers, At-Risk, etc.
Interactive Dashboard	Streamlit dashboard with real-time filters by date, channel, device, and customer segment

---

🚀 Quick Start

Prerequisites

• Python 3.9+
• pip

Installation

# Clone the repository
git clone https://github.com/Yash-Patil-1/UserInsights.git
cd UserInsights

# Create a virtual environment (recommended)
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Generate Data

python src/data_generator.py

This creates data/ecommerce_events.csv with ~34,500 events across 10,000 users over a 6-month period.

Run EDA

python src/explore.py

Generates 7 visualization PNGs in reports/eda_plots/.

Run Cohort & Funnel Analysis

python src/cohorts.py
python src/segments.py

Creates interactive HTML reports in reports/.

Launch Dashboard

streamlit run dashboard.py

Opens the interactive dashboard in your browser.

---

🎛️ Dashboard

The Streamlit dashboard provides real-time analytics with:

Sidebar Filters:

• 📅 Date range picker
• 📡 Channel selector (Organic Search, Paid Search, Social Media, etc.)
• 📱 Device selector (Desktop, Mobile, Tablet)
• 👥 Customer segment filter (Champions, Loyal Customers, etc.)

Visualization Panels:

• KPI Cards — Active Users, Sessions, Purchases, Revenue, Purchase Rate
• Conversion Funnel — Interactive plotly funnel chart with step-by-step drop-off
• Daily Trends — Dual-axis chart (events bar + revenue line)
• Retention Cohorts — Monthly retention heatmap (YlOrRd)
• Channel Performance — Grouped bar chart with conversion rate annotations
• Hourly Activity — Day-of-week × hour heatmap
• Device Breakdown — Donut chart by revenue
• Top Categories — Revenue by product category
• RFM Segments — Segment distribution bar + RFM scatter plot
• Raw Data — Expandable data table preview

---

📂 Project Structure

UserInsights/
├── dashboard.py              # Streamlit interactive dashboard
├── requirements.txt          # Python dependencies
├── .gitignore                # Git ignore rules
├── LICENSE                   # MIT license
├── README.md                 # This file
│
├── src/
│   ├── config.py             # Configuration constants
│   ├── data_generator.py     # Synthetic e-commerce data generator
│   ├── explore.py            # EDA with matplotlib/seaborn plots
│   ├── cohorts.py            # Monthly retention & revenue cohort analysis
│   └── segments.py           # Funnel analysis & RFM customer segmentation
│
├── data/
│   ├── ecommerce_events.csv  # Generated dataset (34K+ events)
│   └── rfm_segments.csv      # Per-user RFM scores & segment labels
│
├── reports/                  # Generated outputs (gitignored)
│   ├── eda_plots/            # 7 PNG EDA charts
│   ├── retention_cohorts.html
│   ├── revenue_cohorts.html
│   ├── retention_curves.html
│   ├── funnel_bar.html
│   ├── funnel_sankey.html
│   ├── funnel_by_channel.html
│   ├── rfm_segments.html
│   └── rfm_scatter.html
│
└── docs/
    ├── getting_started.md    # Setup & quick demo guide
    ├── usage.md              # CLI & dashboard usage reference
    ├── architecture.md       # Pipeline & data model documentation
    └── development.md        # Contributing & development guide

---

🔍 Key Findings

Metric	Value
Overall Conversion Rate	2.8% (visit → purchase)
Month-1 Retention	18.7% average (best cohort: Oct 2025 at 32.4%)
Best Channel	Paid Search (3.2% conversion)
Worst Channel	Social Media (2.4% conversion)
Loyal Customers	151 users driving 45.3% of revenue
Browsing-Only Users	94% of users (9,400 of 10,000)
Avg Order Value	Rs. 38,977
Total Revenue (6 months)	Rs. 28.1M

Funnel Drop-off Points

Step Transition	Drop-off Rate
Visit → View Product	59.9%
View Product → Add to Cart	80.4%
Add to Cart → Checkout	48.7%
Checkout → Purchase	30.5%

The biggest drop-off is View Product → Add to Cart (80.4%), suggesting pricing or product page improvements could have the largest impact on conversion.

---

📚 Documentation

• Getting Started — Prerequisites, installation, quick demo
• Usage Guide — Full CLI reference, dashboard guide, examples
• Architecture — Pipeline stages, data models, design decisions
• Development — Dev setup, testing, adding features, contributing

---

🛠️ Tech Stack

Category	Tools
Data Processing	pandas, numpy
Visualization	matplotlib, seaborn, plotly
Dashboard	streamlit
Synthetic Data	numpy (zipf), random
Statistical Analysis	scikit-learn (quintile scoring)

---

Built by Yash Patil — Data Analyst