Interactive Data Viz

A single repository that combines a reproducible data cleaning pipeline, interactive analytics, and algorithm visualizations.

Problem

Data projects often lack a reproducible path from raw inputs to trustworthy visuals.

Solution

An ETL-style cleaning pipeline with data contracts and reporting, plus interactive plots and algorithm visualizers that make results inspectable.

Tech

Python, pandas, NumPy, matplotlib, mplcursors, pytest, GitHub Actions.

Impact

• Reproducible cleaning with validation
• Interactive analysis with cursor inspection
• Visual explanation of algorithm behavior
• Static dashboard export for public portfolio review

Engineering Focus

• Data contracts and validation before visualization
• Separated modules for ETL, analytics, and algorithms
• Test coverage for both pipeline and algorithm logic

Proof of Work

• The cleaning pipeline validates synthetic input data before charts are generated.
• Tests cover data-cleaning behavior and algorithm logic.
• The exported dashboard screenshot provides a stable review artifact for GitHub.
• Interactive charts use cursor inspection so values can be checked instead of only viewed visually.

Difficult Parts / Tradeoffs

• Interactive charts are useful locally, but GitHub needs static screenshots; the project supports both.
• The dataset is synthetic to avoid privacy issues, so the focus is on process quality rather than business claims.
• Validation is kept explicit before visualization because clean-looking charts can still hide bad data.

Quick Start

python -m venv .venv
.venv\Scripts\python -m pip install -U pip
.venv\Scripts\python -m pip install -e .[dev]
.venv\Scripts\python -m interactive_data_viz

Modules

• cleaning_pipeline/: load, validate, clean, report
• viz/: interactive plots, dashboard export, and algorithm visualizations
• algorithms/: sorting + BST + metrics

Demo Flow

1. Run the app to generate the cleaning report and open the interactive charts.
2. Explore a chart point with the cursor to show exact values.
3. Open the algorithm visualizer and compare metrics.
4. Review docs/images/viz_demo.png as a stable GitHub screenshot of the data-quality flow.

Tests

.venv\Scripts\python -m pytest

Design Decisions

See docs/DECISIONS.md.

Data

• data/demo_students.csv contains synthetic demo records (no real personal data).

Engineering Notes

See docs/ENGINEERING_NOTES.md for architecture rationale, tradeoffs, and project evolution notes.