calcite 0.4.5.1

A desktop application for data analysis and publication-quality graphing.

Calcite

Calcite is a desktop application designed for scientists, researchers, and students who need to perform data analysis and create publication-quality graphs without writing code. It provides a seamless workflow from data import to final plot export, all within a single, user-friendly interface.

日本語のREADMEはこちら (Japanese README here)

✨ Features

Intuitive Data Handling

• Versatile Import: Import data from CSV files or paste directly from spreadsheets (e.g., Excel) via the clipboard.
• Python Integration: Launch the application seamlessly from existing analysis environments by passing a pandas.DataFrame as a direct argument.
• Interactive Table:
  • Sort data in ascending/descending order with a single click or edit column names with a double click.
  • Export the current state of the data (after filtering or sorting) to a new CSV file.
• Advanced Data Manipulation:
  • Reshaping: Easily convert data between wide and long formats using a graphical interface.
  • Filtering: A powerful and advanced filtering tool allows for combining multiple conditions using AND/OR logic.
  • Column Calculation: Dynamically create new columns using formulas like 'ColumnA' * 100.

Publication-Quality Graphing

• Variety of Plot Types: Supports a wide range of plots, including Scatter, Bar, Box, Violin, Point, Line, and Paired Scatter plots.
• Extensive Customization:
  • Fine-tune every aspect of your plot from the GUI, including colors, markers, line styles, font sizes, axis ranges, and log scales.
  • Apply a "Prism-style" aesthetic by removing the top and right spines of the graph.
  • Overlay individual data points on summary plots like bar charts and box plots.

Comprehensive Statistical Analysis

• Basic Tests: Independent & Paired t-tests, Mann-Whitney U, Wilcoxon signed-rank.
• Group Comparisons: One-way ANOVA & Kruskal-Wallis with post-hoc tests (Tukey, Dunn).
• Regression: Linear and non-linear (4-parameter logistic, 4PL) regression, with R² values displayed on the graph.
• Correlations & Associations: Spearman's correlation and Chi-squared tests.
• Automatic Annotations: Automatically adds statistical significance (*) to your plots based on the robust logic of the statannotations library.

e.g.

High-Resolution Export

• Save your graphs as PNG, JPEG, SVG, or PDF at 300 DPI, ready for any publication or presentation.

🛠️ Installation

This project is currently under development. The installation method is as follows. Python 3.10 or higher is required.

pip install calcite

🚀 Quick Start

1. Launch Calcite from your terminal:

   calcite
   

   or

   import pandas as pd
   from calcite.main import plot
   
   data = {
       'Category': ['A', 'A', 'B', 'B'],
       'Value': [10, 12, 15, 17]
   }
   df = pd.DataFrame(data)
   # -----------------------------
   
   plot(data=df)
   

2. Import data using File > Open CSV... or paste from your clipboard using Edit > Paste.

   • 💡 Tidy Data format (=Long-form) is recommended
   • Calcite is designed around the principles of Tidy Data. This is a data structure where:
     • Each variable forms a column (e.g., "Genotype", "Concentration", "Measurement").
     • Each observation forms a row.
     • Each type of observational unit forms a table.
   • This format is the most suitable for statistical analysis and graphing on a computer. If your data is in a "wide" format (e.g., separate columns for Control Group, Drug A Group, etc.), you can easily convert it to Tidy Data using Calcite's Data > Restructure (Wide to Long)... feature.

   Tidy data (Ref. Seaborn) (https://seaborn.pydata.org/tutorial/data_structure.html)

3. Select a graph type from the toolbar (e.g., Scatter Plot, Bar Chart).

4. In the "Data" tab at the bottom right, select the columns for the X and Y axes.

5. Customize the graph's appearance using the "Format," "Text & Legend," and "Axis" tabs.

6. Perform statistical analysis from the "Analysis" menu.

7. Save your graph using File > Save Graph As....

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.