Skip to content

Latest commit

 

History

History
144 lines (108 loc) · 6.1 KB

README.md

File metadata and controls

144 lines (108 loc) · 6.1 KB

Probability-Spreadsheet: A Probabilistic Spreadsheet

A simple proof-of-concept probabilistic spreadsheet. Python recreation of Montesheet

Probability-Spreadsheet is an experimental spreadsheet application built with PyQt5 and Python, allowing you to seamlessly mix probability distributions with classical spreadsheet-style arithmetic. It is a Python recreation of Montesheet See also Probabilistic Spreadsheet. This project is NOT affiliated with Montesheet.

Screenshots

image

image

Features

  • Basic Spreadsheet Operations

    • Type values or formulas into cells (e.g., =A1 + B2)
    • Cell references: refer to a cell by its column letter + row number (A1, B3, etc.)
    • Standard arithmetic and referencing of other cells

  • Probabilistic Distributions

    • Use built-in distribution functions like normal, uniform, exponential, and poisson
    • Combine them with normal arithmetic: =normal(10,2) + uniform(0,5)
    • Each distribution cell automatically displays a small sparkline of its distribution
    • Double-click a distribution cell to see a detailed histogram plot with summary statistics

  • Monte Carlo Engine

    • Under the hood, each distribution is sampled (e.g., 5,000 times by default)
    • The distribution’s mean and sparkline are shown in the cell header
    • You can recalculate the entire spreadsheet to regenerate new random samples
    • Composite cells referencing distributions will also update

  • Dependency Tracking

    • Changing any cell (or distribution) automatically invalidates dependent cells
    • The spreadsheet re-evaluates only what is necessary
    • On “Calculate,” the entire network of cells is re-sampled and re-evaluated

  • Automatic Sizing

    • Cells containing distributions are automatically resized (larger height and width)
    • Regular numeric or text cells remain in their default size
    • The row or column containing at least one distribution stays large, even if other cells in that row or column are numeric

  • Load/Save via XML

    • Save your spreadsheet to an XML file (with File → Save)
    • Load a previously saved spreadsheet via File → Load
    • The distribution formulas (e.g. =normal(10,2)) and references are all restored from the XML

  • Decimal Precision

    • A small text box in the toolbar lets you specify the number of decimals to display (default is 3)
    • Press the FIX button to apply the new precision to all cells

Installation & Usage

  1. Install Requirements

    • Python 3.x
    • pip install pyqt5 matplotlib numpy

  2. Run the Script

    python probability-spreadsheet.py

    A PyQt5 window should appear with 20 rows × 10 columns, a formula bar, and a toolbar.

  3. Entering Data

    • Click a cell to select it. In the formula bar at the top, type either:
      • A plain number, e.g. 123.45
      • A formula referencing another cell, e.g. =A1 + 2.5
      • A distribution formula, e.g. =normal(10,2,0,20) (normal distribution truncated between 0 and 20)
    • Press Enter or click away to set the cell.

  4. Recalculation

    • Click Calculate on the toolbar to force the entire spreadsheet to re-sample distributions and update dependent cells.

  5. Double-Click on Distribution Cells

    • Any cell showing a sparkline (which means it contains a distribution) can be double-clicked
    • A dialog opens, displaying a histogram and key statistics (min, 10th percentile, mean, 90th percentile, max)

  6. Saving and Loading

    • Save: Choose “Save” in the toolbar, select an XML file location
    • Load: Choose “Load” in the toolbar, pick an XML file that was previously saved. The script will restore your spreadsheet’s expressions.

Built-in Distribution Functions

You can use the following distribution functions inside your spreadsheet formulas:

  • normal(mean, std, min, max)
    • Creates a normal distribution with given mean and std.
    • Optional min and max parameters let you truncate the distribution.
  • uniform(min, max)
    • Uniform distribution in [min, max].
  • exponential(rate)
    • Exponential distribution with the given rate (mean = 1/rate).
  • poisson(mean)
    • Poisson distribution with given mean.
  • mmax(x, y)
    • Takes the element-wise maximum of distributions (or numbers) x and y.
  • choose(x, y)
    • Randomly picks x or y (each with 50% chance) at each sample.

Additionally, you can call:

  • mean(x) — returns the minimum, mean, or maximum of distribution x as a single scalar.
  • min(x)
  • max(x)
  • perc(x, p) — returns the p-th percentile (0–100) of distribution x.

Example Calculation

Here’s a simple demonstration:

  1. In A1, type:

    =normal(10, 2)

    This cell will contain a Normal(μ=10, σ=2) distribution.

  2. In B1, type:

    =uniform(0, 5)

    This cell will contain a Uniform(0, 5) distribution.

  3. In C1, type:

    =A1 + B1

    This will be the sum of the Normal and Uniform distributions. A sparkline will appear showing the distribution of the sum.

  4. Click Calculate to resample. Cells A1, B1, and C1 will update to show new sparkline shapes and mean values.

    • If you double-click C1, you’ll see a histogram of the sum.

  5. Try referencing a single statistic in another cell. For instance, in D1, type:

    =mean(C1)

    This will convert the distribution in C1 into a single numeric mean value.

  6. Click Calculate again, and you’ll see D1 update with the new mean of the sum distribution.

Contributing

If you find a bug or want to add features, feel free to open issues or submit pull requests on GitHub.

License

This project is provided “as-is” under the MIT License. Feel free to use and modify it as you wish.