Smartphone Sales Analytics

An end-to-end data-analyst project on FY2025 smartphone sales in the Indian market. A single CSV feeds a validating pandas pipeline, a DuckDB/SQL query layer, a Jupyter EDA notebook, and an interactive Streamlit dashboard that ranks the top-performing products across the Budget, Mid-Range, Premium, and Flagship price segments.

🔗 Live demo: https://smartphone-sales-analytics.streamlit.app

Built with: Python · pandas · DuckDB (SQL) · Plotly · Streamlit · Jupyter

Dataset

data/smartphones.csv — 60 products with price, units sold, rating, quarterly sales, and key specs. Price segments: Budget (<=₹15K), Mid-Range (₹15–30K), Premium (₹30–50K), Flagship (>₹50K).

Setup

python3 -m venv .venv && . .venv/bin/activate
pip install -r requirements.txt          # runtime (dashboard + SQL)
pip install -r requirements-dev.txt      # adds tests + notebook tooling

Usage

• Dashboard: streamlit run app.py
• Build the DuckDB database: python db/build_db.py (writes db/smartphones.db)
• Run SQL: duckdb db/smartphones.db < sql/queries.sql (requires the DuckDB CLI)
• EDA notebook: jupyter notebook notebooks/analysis.ipynb (needs requirements-dev.txt)
• Tests: pytest (needs requirements-dev.txt)

Deployment

The dashboard is deployed on Streamlit Community Cloud: push to master, then point a new app at app.py — Cloud installs requirements.txt and serves it. Every push to master redeploys automatically.

Layout

Path	Purpose
data/smartphones.csv	Single source of truth
src/config.py	Segment bins and color config
src/data_loader.py	Load, validate, derive (segment, revenue, perf_score)
src/analysis.py	Shared aggregation functions
db/build_db.py, sql/queries.sql	DuckDB + reference SQL
notebooks/analysis.ipynb	Exploratory analysis
app.py	Streamlit dashboard
tests/	pytest suite