---

NOS-TLPlot is an open-source Python tool for visualising Newcastle–Ottawa Scale (NOS) risk-of-bias assessments.
It converts NOS star ratings into publication-ready traffic-light plots and 10 specialized visualizations, enabling reviewers and readers to interpret study-level risk-of-bias results clearly and reproducibly.

📂 Code & Archive (Zenodo DOI): 10.5281/zenodo.17065214

📃 Software Metapaper (Journal of Open Research Software DOI): 10.5334/jors.635

---

✨ Key Features

• 12 Visualization Types: Traffic-light plots, radar charts, heatmaps, dot profiles, donut charts, lollipop charts, line plots, and more
• Publication-Quality Output: Export figures in .png, .pdf, .svg, .eps formats
• Multiple Themes: Traffic-light and grayscale themes for various publication requirements
• Interactive Web App: Built with Streamlit for simple data upload, preview, and figure export
• Command Line Interface: For batch processing and automated figure generation
• Domain-Specific Risk Assessment: Automatically converts NOS stars (0–9) to categorical RoB (Low/Moderate/High)
• Customizable Output: Adjustable figure sizes, line thickness, and color themes
• Scope-Limited: Designed exclusively for NOS evaluations of non-randomised studies

---

USE

• Vercel web leading to Streamlit → nos-tlplot.vercel.app

Vercel user-interface

---

• To directly use it → nos-tlplot.streamlit.app

Streamlit user-interface

---

📥 Installation (Optional)

Prerequisites

• This code is based on Python-3.12.3 (however works with 3.11+)
• pip package manager

Installation Steps

# Clone the repository
git clone https://github.com/aurumz-rgb/NOS-TLPlot.git
cd NOS-TLPlot

# Install dependencies
pip install -r requirements.txt

---

⚡ Usage (directly via Streamlit web OR)

1️⃣ Streamlit Web App (Recommended)

cd NOS-TLPlot
streamlit run app.py

Features:

• Upload CSV/Excel files
• Real-time visualization preview
• Choose from 12 plot types
• Download publication-ready figures in multiple formats
• Theme switcher (Traffic-light / Grayscale)

Steps:

1. Run the above command
2. Open the local Streamlit URL (default: http://localhost:8501)
3. Upload your NOS dataset
4. Choose visualization and theme
5. Preview and download figures

---

2️⃣ Python Script (Command Line) (For Reviewers to ensure reproducibility and Transparency)

cd NOS-TLPlot
# Basic usage
python3 nos_tlplot.py sample.csv output.png

# With theme
python3 nos_tlplot.py sample.csv output-traffic-light.png gray

---

🧩 Python Sample

---

Parameters:

Parameter	Description
input_file	Path to CSV/Excel file containing NOS data
output_file	Output file name and extension
theme	Optional theme: traffic_light (default) or gray

Generated Outputs:

File	Description
_traffic-light.png	Classic traffic-light bubble plot
_radar.png	Radar chart of domain scores
_heatmap.png	Color heatmap for domain-level bias
_dot_profile.png	Dot-style bias visualization
_table.png	Summary table with domain scores
_donut.png	Donut chart of overall bias levels
_line_ordered.png	Sequential line plot of domain bias
_lollipop.png	Lollipop chart for comparative bias
_pie.png	Proportional risk-of-bias pie
_stacked_area.png	Stacked area visualization over domains
_star_dist.png	Star distribution visualization

---

Input Data Format

Your file should have these columns:

Column Name	Description	Valid Range
Author, Year	Study identifier	Text
Representativeness	Domain 1	0–1
Non-exposed Selection	Domain 2	0–1
Exposure Ascertainment	Domain 3	0–1
Outcome Absent at Start	Domain 4	0–1
Comparability (Age/Gender)	Domain 5	0–2
Comparability (Other)	Domain 6	0–2
Outcome Assessment	Domain 7	0–1
Follow-up Length	Domain 8	0–1
Follow-up Adequacy	Domain 9	0–1
Total Score	Sum of stars	0–9
Overall RoB	Risk of bias	Low / Moderate / High

💡 Tip: Always include your raw NOS scoring table in supplementary materials for reproducibility.

---

🎨 Visualization Types

1. Traffic-light bubble Plot – Standard bubble risk-of-bias visualization.
2. Radar Chart – Displays study performance across domains.
3. Heatmap – Visual overview of domain-level variation.
4. Dot Profile – Shows domain-level bias in compact form.
5. Donut Chart – Visualizes proportions of bias categories.
6. Lollipop Plot – Combines numerical and categorical domains.
7. Stacked Area Chart – Displays temporal or comparative changes.
8. Pie Chart – Quick overview of overall bias distribution.
9. Line Ordered Plot – Connects domain bias levels for each study.
10. Table View – Tabular representation of bias domains.
11. Radar (Thematic) – Theme-adapted radar chart (gray/colored).
12. Star distribution – Star-adapted plot (only traffic-light).

Note: Radar Plots and Dot Profile plot is only limited to 5 studies.

---

NOS Scoring → Risk-of-Bias Conversion

Total Stars	Interpretation	Risk Category
7–9	High-quality study	Low RoB
4–6	Moderate-quality study	Moderate RoB
0–3	Poor-quality study	High RoB

Conversion follows Newcastle–Ottawa Scale standards for cohort/case-control/cross-sectional designs.

---

🔧 Technical Details

• Core Engine: matplotlib, seaborn, numpy, pandas
• Web UI: streamlit
• Table Rendering: matplotlib.table
• Plot Layout Management: GridSpec
• Color Systems: Custom mcolors, traffic-light mapping
• Data Handling: CSV/Excel file input with automatic parsing
• Batch Plotting: Parallel generation for multiple figure types
• Export Quality: 300 DPI (default), publication-ready vector output

---

🧩 Repository Structure

NOS-TLPlot/
├── app.py                 # Streamlit web app
├── nos_tlplot.py          # Main plotting engine
├── requirements.txt       # Dependencies
├── README.md              # Project documentation
├── citation.cff           # Citation metadata
├── LICENSE                # Apache 2.0 License
├── examples               # All the Sample outputs

---

Support

• For usage questions, open a Discussion
• For bug reports or feature requests, open an Issue
• Email: mail

---

📄 Acknowledgment:

I sincerely thank the Journal of Open Research Software (JORS) for providing a full publication waiver supporting this software.

---

🎯 Citation

NOS-TLPlot Software code is cited:

Sahu, V. (2025). NOS-TLPlot: Visualization Tool for Newcastle–Ottawa Scale in Meta-Analysis (v2.0.3). Zenodo. DOI: 10.5281/zenodo.17065214

NOS-TLPlot Software Metapaper is cited:

Sahu, V. (2026). NOS-TLPlot: A Specialized Python Tool for Visualizing Newcastle–Ottawa Scale Risk-of-Bias Assessments. Journal of Open Research Software, 14(1), 7. DOI: 10.5334/jors.635

@software{Sahu2025,
  author = {Sahu, Vihaan},
  title = {NOS-TLPlot: Visualization Tool for Newcastle–Ottawa Scale in Meta-Analysis (v2.0.3)},
  year = {2025},
  doi = {10.5281/zenodo.17065214},
  url = {https://doi.org/10.5281/zenodo.17065214},
  version = {2.0.3}
}

@article{Sahu2026,
  author = {Sahu, Vihaan},
  title = {NOS-TLPlot: A Specialized Python Tool for Visualizing Newcastle–Ottawa Scale Risk-of-Bias Assessments},
  journal = {Journal of Open Research Software},
  volume = {14},
  number = {1},
  pages = {7},
  year = {2026},
  doi = {10.5334/jors.635},
  url = {https://doi.org/10.5334/jors.635}
}

---

Sample Plots

All the Output Sample plots

NOS star plot

NOS bubble plot

Domain Scores Radar Chart by Study

Theme-based Domain Scores Radar Chart

Domain Scores Ordered by Total Score

Total NOS Scores by Study (Lollipop Chart)

Domain Score Profiles by Study

Risk Distribution by Domain (Stacked Area Chart)

Risk Donut Distribution by Domain

Distribution of Overall Risk of Bias Pie

Risk of Bias by Domain and Study by heatmap

NOS Scores by Study