Statistical Analysis: Ping Pong Precision Model

---

Project Overview

This repository contains a statistical analysis case study that evaluates ping pong serve precision through a reproducible R pipeline and a lightweight web dashboard. The goal is to move from raw experimental observations to validated statistical outputs, exported artifacts, and a browser-based summary that makes the results easier to review.

The public version focuses on:

• reproducible statistical processing in R
• exported JSON and PNG outputs for dashboard delivery
• inferential tests that validate model fit and timing behavior
• a clear analytical summary for technical and non-technical review

---

Challenge / Solution / Impact

Challenge	Solution	Impact
Validate whether serve attempts fit a defendable statistical model	Built an R pipeline with descriptive and inferential analysis, including goodness-of-fit testing	The Negative Binomial fit was accepted with p = 0.6603
Manual analysis made results harder to reproduce and communicate	Automated JSON and PNG exports from the analysis script	The dashboard can load consistent outputs without rewriting results by hand
Static notebooks and raw tables were harder to scan quickly	Delivered a lightweight dashboard that presents charts and key metrics in one place	Statistical findings became easier to review and share

---

Key Metrics

Metric	Value	Interpretation
Sample size	309 observations	Experimental dataset used for the full analysis
Chi-square p-value	0.6603	Goodness-of-fit accepted for the selected model
Mean serve time	1.945s	Average time stayed below the 2-second threshold
Right-handed share	64.4%	Population split is documented in the exported results
Correlation	0.0494	Weak relationship between attempts and serve time

---

Pipeline Architecture

Raw CSV data
     ->
R analysis script (scripts/ping_pong_analysis.R)
     ->
JSON metrics + PNG charts
     ->
Web dashboard summary

Layer	Technologies	Output
Data input	CSV experiment data	Raw observations
Analysis	R, tidyverse, ggplot2, jsonlite	Statistical results and visualizations
Delivery	HTML, CSS, JavaScript	Interactive dashboard that reads exported outputs

---

Methodology Snapshot

The analysis workflow includes:

• descriptive statistics for serve time and attempts
• chi-square goodness-of-fit validation
• one-sample t-test against the 2-second threshold
• correlation analysis between attempts and serve time
• exported charts and JSON summaries for dashboard delivery

---

Tech Stack

Layer	Technologies
Data processing	R, tidyverse, ggplot2, jsonlite
Visualization	ggplot2, ggsave
Frontend	HTML5, CSS3, JavaScript
Hosting	GitHub Pages

---

Quick Start

# clone repository
git clone https://github.com/Sam-24-dev/Analisis-Ping-Pong.git
cd Analisis-Ping-Pong

# install required R packages
R -e "install.packages(c('tidyverse','jsonlite','here'))"

# run the analysis pipeline
Rscript scripts/ping_pong_analysis.R

# or use the existing Makefile
make pipeline
make serve

---

Project Structure

Analisis-Ping-Pong/
|- data/
|  |- raw/                         # Original experiment files
|  `- ping_pong_experiment_data.csv
|- docs/                           # Technical architecture notes
|- scripts/                        # R analysis pipeline
|- web/
|  |- images/                      # Exported PNG charts
|  |- results/                     # JSON statistical outputs
|  |- index.html                   # Dashboard entrypoint
|  |- script.js
|  `- styles.css
|- Makefile
|- LICENSE
`- README.md

---

Documentation

• Technical architecture: docs/architecture.md
• Main analysis script: scripts/ping_pong_analysis.R
• Exported results: web/results/analysis_results.json

---

Author

Samir Caizapasto
Junior Data Engineer & Analyst

---

If this analysis was useful or interesting, consider starring the repository.