Install & Compatibility
Where this runs
tested against v0.5.0.1 · pip install
no network on importno background threads
Install × environment matrix
Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux
py 3.10
✕ build_error
✓ 24.69s
py 3.11
✕ build_error
✓ 22.58s
py 3.12
✕ build_error
4/8 runs
py 3.13
✕ build_error
4/8 runs
py 3.9
✕ build_error
✓ 27.6s
556MB installed
● package 556MB
Code
Verified usage
Verified import paths — ran on the pinned version, not inferred.
pygwalker
✓ import pygwalker as pyg
StreamlitRenderer
✓ from pygwalker.api.streamlit import StreamlitRenderer
Required for integrating Pygwalker into Streamlit applications.
Connector
✓ from pygwalker.data_parsers.database_parser import Connector
Used for connecting to database sources like Snowflake directly.
This quickstart demonstrates how to initialize Pygwalker with a pandas DataFrame, launching an interactive UI in your Jupyter Notebook or compatible environment. The `pyg.walk()` function takes a DataFrame and opens a graphical interface for drag-and-drop data exploration. For larger datasets, consider adding `kernel_computation=True` to leverage DuckDB for enhanced performance.
import pandas as pd
import pygwalker as pyg
# Create a sample DataFrame
data = {
'City': ['New York', 'Los Angeles', 'Chicago', 'Houston', 'Phoenix'],
'Population': [8804190, 3898747, 2716075, 2325502, 1680992],
'Area_sq_mi': [302.6, 468.7, 227.6, 637.4, 517.6]
}
df = pd.DataFrame(data)
# Launch Pygwalker UI
# You can also use pyg.walk(df, kernel_computation=True) for better performance on large datasets.
walker = pyg.walk(df)
Debug
Known issues
deprecatedThe `use_kernel_calc` parameter in `pyg.walk()` is deprecated.fixUse `kernel_computation=True` instead for enabling the DuckDB-powered computation engine.
affects: <=0.4.x
breakingAPI changes were introduced in version 0.4.8, affecting general and Streamlit APIs.fixSpecifically, the `dark` parameter was renamed to `appearance`. Streamlit APIs shifted to a more declarative style. Refer to the official GitHub changelog or documentation for specific migration details related to your usage patterns.
affects: 0.4.8 and later
gotchaWhen using the `spec` parameter to save chart configurations, changes made in the UI are not automatically saved.fixUsers must manually click the 'Save' button within the Pygwalker UI to persist chart metadata to the specified JSON file or string. An 'autosave' feature is planned for future releases.
affects: All versions up to 0.5.0.1
gotchaSome cloud-integrated features, such as saving charts to the cloud or using advanced GPT-powered features, require a Kanaries token.fixObtain a Kanaries token from `kanaries.net` and configure it (e.g., via `pygwalker config --set kanaries_token=YOUR_TOKEN`) to enable these functionalities.
affects: All versions supporting cloud features
Errors
Common errors & fixes
ERROR: Failed building wheel for quickjs
Pygwalker's dependency `quickjs` requires C++ build tools for compilation on Windows, which are often missing from a standard Python installation.
fixInstall the pre-built wheel by running `pip install --only-binary :all: pygwalker` or install Microsoft C++ Build Tools (Desktop development with C++).
ModuleNotFoundError: No module named 'pygwalker'
The `pygwalker` library is not installed in the active Python environment or is not accessible in the deployment environment.
fixInstall the library using `pip install pygwalker` for pip or `conda install -c conda-forge pygwalker` for Conda.
AttributeError: 'DataFrame' object has no attribute 'walk'
The `walk` function is a top-level function in the `pygwalker` library (typically aliased as `pyg`), not a method of a pandas DataFrame object.
fixImport `pygwalker` (conventionally as `pyg`) and call `pyg.walk()` directly with your DataFrame: `import pygwalker as pyg; pyg.walk(df)`.
(Descriptive) Pygwalker freezes or encounters memory errors with large datasets.
By default, `pygwalker` might process data in-memory, which can lead to performance issues or crashes when working with very large DataFrames (e.g., >500MB).
fixEnable the `kernel_computation` parameter to leverage Pygwalker's DuckDB-based engine for efficient processing of larger datasets: `pyg.walk(df, kernel_computation=True)`.
Upgrade
Version history
0.5.0.1latest on PyPI · released Apr 4, 2026
Audit
Dependencies
pandasrequiredCore data structure for `pyg.walk()` input.
ipywidgetsrequiredProvides the interactive UI within Jupyter environments.
duckdboptionalUsed as the computation engine when `kernel_computation=True` for improved performance with larger datasets.
streamlitoptionalRequired for building web applications with Streamlit using `StreamlitRenderer`.
snowflake-connector-pythonoptionalRequired for connecting to Snowflake when using the `snowflake` extra.