Inspire the best visualization for your research data | 激发科研数据最佳可视化方案
👉 https://scivizkit-hvbzujsahst6uec2pupvrx.streamlit.app/
No installation needed — upload your data and explore 80+ chart types instantly. 无需安装 — 直接上传数据，即刻探索 80+ 种科研图表。
SciVizKit is an open-source scientific visualization toolkit built with Streamlit. Upload any dataset (CSV or Excel), and it automatically generates 80+ chart types across 10 categories — from standard statistics charts to domain-specific scientific plots. Pick the best one for your paper, download publication-quality figures (300 DPI), and copy the Python code.
- 📊 80+ chart types across 10 categories (Distribution, Comparison, Correlation, Time Series, Proportional, Network, Scientific, Text, Geographic, 3D)
- 🌳 Chart Guide — interactive decision tree that recommends the best chart based on your data and goal
- 🖼️ Figure Panel Builder — combine multiple charts into a single publication-ready multi-panel figure (PNG/SVG, 300–600 DPI)
- 🎨 Journal color palettes — Nature, Science, Cell, ACS, Colorblind Safe
- 🤖 Smart column detection — auto-detects numeric, categorical, datetime columns
- 🎯 Domain-aware — Biology & Genomics, Chemistry & Materials, Medicine & Clinical, Physics & Engineering, Social Science
- 💻 Copy-paste Python code for every chart
- ⚡ Lazy loading — generate charts by category, not all at once
- ⭐ Favorites — star charts you like, filter to favorites only
- 🚀 Zero-config — works with any tabular data, no setup needed
| Category | Count | Examples |
|---|---|---|
| Distribution | 11 | Histogram, Violin, Ridgeline, Raincloud, ECDF |
| Comparison | 16 | Bar, Lollipop, Dumbbell, Slope, Tornado, Radial Bar, Bar + Significance ✨ |
| Correlation | 9 | Scatter, Hexbin, 2D KDE, Pair Plot, Parallel Coords |
| Time Series | 9 | Line, Streamgraph, Bump Chart, Calendar Heatmap, Candlestick |
| Proportional | 9 | Treemap, Sunburst, Waffle, Marimekko, Circle Packing, Jade Ring (玉珏图) ✨ |
| Network | 6 | Sankey, Chord, Arc Diagram, Alluvial, Network Graph |
| Scientific | 16 | Volcano, PCA, UMAP, t-SNE, ROC, Kaplan-Meier, Manhattan, Forest, Radial Bar + Sig ✨ |
| Text | 2 | Word Cloud, Venn Diagram |
| Geographic | 2 | Choropleth Map, Bubble Map |
| 3D | 3 | 3D Scatter, 3D Surface, 3D Bar |
| Total | 83 |
✨ = newly added, NGplot-inspired charts
SciVizKit gives you open-source Python code you can copy and own. For rapid interactive exploration with 700+ ready-made templates (no coding required), NGplot is an excellent companion:
| Feature | SciVizKit | NGplot |
|---|---|---|
| Open-source | ✅ | ❌ |
| Python code export | ✅ | ✅ |
| Chart templates | 83 | 700+ |
| No-code web UI | ✅ Streamlit | ✅ |
| Error bar + significance | ✅ (bar_sig, radial_bar_sig) | ✅ |
| 玉珏图 / Jade Ring | ✅ | ✅ |
| Publication-quality export | ✅ 300 DPI | ✅ |
| Local/offline | ✅ | ❌ |
SciVizKit charts inspired by NGplot are marked ✨ in the table above.
git clone https://github.com/Yang1Bai/SciVizKit.git
cd SciVizKit
pip install -r requirements.txt
streamlit run app.pyOpens at http://localhost:8501 🎉
- Fork this repo on GitHub
- Go to share.streamlit.io and sign in with GitHub
- Click "New app" → select
SciVizKit→ branchmain→ fileapp.py - Click Deploy — live in ~3 minutes
SciVizKit/
├── app.py # Main Streamlit app
├── requirements.txt
├── .streamlit/config.toml # Theme config
├── src/
│ ├── data_analyzer.py # Column type detection
│ ├── chart_registry.py # Registry of 80+ chart types
│ ├── decision_tree.py # Chart recommendation tree
│ ├── figure_panel.py # Multi-panel figure builder
│ ├── themes/palettes.py # Journal color palettes
│ └── generators/ # Chart generation modules
│ ├── distribution.py
│ ├── comparison.py
│ ├── correlation.py
│ ├── timeseries.py
│ ├── proportional.py
│ ├── network.py
│ ├── scientific.py
│ ├── text_viz.py
│ ├── geo_viz.py
│ └── threed_viz.py
└── examples/
├── sample_general.csv
├── sample_biology.csv
└── sample_chemistry.csv
- Fork → create branch
feature/your-chart - Add chart metadata to
src/chart_registry.py - Implement generator in the appropriate
src/generators/*.py - Submit a pull request
MIT License — see LICENSE for details.
SciVizKit 是一个基于 Streamlit 的开源科研可视化工具。上传任意数据集（CSV 或 Excel），自动生成 80+ 种图表，覆盖 10 大类型——从基础统计图到领域专用科研图。选出最适合论文的方案，下载发表级图片（300 DPI），并一键复制 Python 代码。
- 📊 80+ 种图表，覆盖 10 大类（分布/比较/相关性/时间序列/比例/网络/科研专用/文本/地理/3D）
- 🌳 图表引导 — 交互式决策树，根据你的数据和目标推荐最佳图表
- 🖼️ 多图面板 — 把多张图拼成一张发表级多面板图（PNG/SVG，300–600 DPI）
- 🎨 期刊配色 — Nature、Science、Cell、ACS、色盲友好
- 🤖 智能列检测 — 自动识别数值、分类、时间列
- 🎯 领域专用 — 生物基因组、化学材料、医学临床、物理工程、社会科学
- 💻 每张图附带可复制的 Python 代码
- ⚡ 懒加载 — 按类别分批生成，秒级响应
- ⭐ 收藏功能 — 标记喜欢的图，一键过滤
- 🚀 零配置 — 任何表格数据直接用，无需额外设置
| 类别 | 数量 | 代表图表 |
|---|---|---|
| 分布 Distribution | 11 | 直方图、提琴图、Ridgeline、雨云图、ECDF |
| 比较 Comparison | 15 | 条形图、棒棒糖、哑铃图、斜坡图、龙卷风图 |
| 相关性 Correlation | 9 | 散点图、六边形密度、2D KDE、平行坐标 |
| 时间序列 Time Series | 9 | 折线图、流图、排名图、日历热力图、K线图 |
| 比例 Proportional | 8 | 树图、旭日图、华夫饼图、马赛克图、圆形填充 |
| 网络 Network | 6 | Sankey、弦图、弧线图、冲积图、网络图 |
| 科研专用 Scientific | 15 | 火山图、PCA、UMAP、t-SNE、ROC、生存曲线、曼哈顿图 |
| 文本 Text | 2 | 词云、韦恩图 |
| 地理 Geographic | 2 | 等值线地图、气泡地图 |
| 3D | 3 | 3D散点、3D曲面、3D柱图 |
| 合计 | 80+ |
git clone https://github.com/Yang1Bai/SciVizKit.git
cd SciVizKit
pip install -r requirements.txt
streamlit run app.py浏览器访问 http://localhost:8501 即可使用 🎉
- Fork 本仓库到你的 GitHub 账号
- 访问 share.streamlit.io，用 GitHub 账号登录
- 点击 "New app" → 选择仓库
SciVizKit→ 分支main→ 入口文件app.py - 点击 Deploy — 约 3 分钟后获得公开链接
SciVizKit/
├── app.py # 主应用
├── requirements.txt # 依赖
├── .streamlit/config.toml # 主题配置
├── src/
│ ├── data_analyzer.py # 数据列类型检测
│ ├── chart_registry.py # 80+ 图表注册表
│ ├── decision_tree.py # 图表推荐决策树
│ ├── figure_panel.py # 多图面板构建器
│ ├── themes/palettes.py # 期刊配色方案
│ └── generators/ # 各类图表生成模块
└── examples/ # 示例数据集
- Fork 本仓库，创建分支
feature/你的图表名 - 在
src/chart_registry.py添加图表元数据 - 在对应的
src/generators/*.py实现生成函数 - 提交 Pull Request
MIT License，详见 LICENSE。
为科研人员、数据科学家和所有热爱好图表的人而生。 Made for researchers, data scientists, and anyone who loves great charts.