Lets-Plot

4.9.0 · active · verified Thu Apr 16

Lets-Plot is an open-source, multiplatform plotting library for statistical data based on the Grammar of Graphics, heavily influenced by R's ggplot2. It provides a Python API for creating both static and interactive charts in notebooks and IDEs, supporting various data formats like Pandas DataFrames and NumPy. The current version is 4.9.0, with a regular release cadence often tied to its Kotlin counterpart.

Common errors

```
ImportError: DLL load failed while importing lets_plot_kotlin_bridge: The specified module could not be found.
```
cause This error typically occurs on Windows systems, especially when using Anaconda or Miniconda, and indicates missing native backend components required by Lets-Plot's Kotlin bridge.

fix Install the `m2w64-toolchain` using `conda install m2w64-toolchain` before installing `lets-plot` with pip, or ensure your environment has the necessary C++ build tools.
```
Plots are not displayed in my Jupyter Notebook cell output.
```
cause The HTML rendering functionality for notebooks needs to be explicitly initialized once per session.

fix Add `from lets_plot import LetsPlot; LetsPlot.setup_html()` to the top of your notebook.
```
Error: geom_contour : error if data is not sorted properly
```
cause The `geom_contour` geometry expects input data to be sorted correctly for proper contour calculation.

fix Ensure your DataFrame is sorted by the x- and y-variables used in `geom_contour` before plotting. For example, `df.sort_values(by=['x_var', 'y_var'], inplace=True)`.

Warnings

breaking The Geocoding API underwent a complete overhaul in Lets-Plot v2.0.0, rendering previous API implementations incompatible.
Fix: Refer to the updated Geocoding documentation for the new API structure and migrate your code accordingly.
breaking Version 4.0.0 introduced significant internal package refactoring. While the Python API was largely unaffected, default qualitative color palettes (now 'Set1' from 'Set2') and default sizes for points and lines changed. It also caused partial incompatibility with older Kotlin API versions.
Fix: Review plots for visual consistency and adjust color scales (`scale_color_manual`) or geom sizes as needed. For multiplatform projects, ensure Kotlin API versions are updated to maintain compatibility.
breaking In version 4.9.0 (primarily affecting Kotlin/JVM), JavaFX artifacts were removed, and the `lets-plot-image-export` module was deprecated. The `DefaultPlotPanelBatik` was replaced by `SwingPlotPanel` for JVM Swing applications.
Fix: For JVM Swing applications, replace `DefaultPlotPanelBatik` with `SwingPlotPanel` from `org.jetbrains.letsPlot.awt.plot.swing`. For image export, use `PlotImageExport` from the `platf-awt` module.
gotcha Forgetting to call `LetsPlot.setup_html()` in Jupyter, JupyterLab, Google Colab, or other notebook environments will result in plots not being rendered or displayed.
Fix: Always include `from lets_plot import LetsPlot; LetsPlot.setup_html()` at the beginning of your notebook session.

Install

pip install lets-plot Standard installation
conda install m2w64-toolchain && pip install lets-plot Windows (Anaconda/Miniconda) with native backend support

Imports

All functions
```
from lets_plot import *
```
The library encourages importing all functions directly for ease of use due to its extensive API.
LetsPlot.setup_html()
wrong:
```
from lets_plot import * # without calling setup_html()
```
correct:
```
from lets_plot import LetsPlot
LetsPlot.setup_html()
```
Crucial for displaying plots in Jupyter notebooks and other HTML environments. Must be called once per session.

Quickstart

This quickstart demonstrates a basic scatter plot with a smoothed regression line, colored by category, using Pandas DataFrame as input. It includes the essential `LetsPlot.setup_html()` call for notebook environments.

import numpy as np
import pandas as pd
from lets_plot import *

LetsPlot.setup_html()

data = pd.DataFrame({
    'x': np.random.normal(0, 1, 100),
    'y': np.random.normal(1, 1.5, 100),
    'category': np.random.choice(['A', 'B'], 100)
})

plot = (ggplot(data, aes(x='x', y='y', color='category')) +
        geom_point(size=5, alpha=0.6) +
        geom_smooth(method='lm', color='black', linetype='dashed') +
        ggtitle('Scatter Plot with Regression Line')
       )

plot

view raw JSON →