Skip to content

Installation

freshdata requires Python ≥ 3.9 and pandas ≥ 1.5.

Basic install

pip install freshdata-cleaner

This installs the pandas + NumPy core plus FreshData's standard reporting and self-contained HTML visualization. You do not need an extra to call fd.clean(df).summary(), fd.clean(df).report(), or fd.clean(df).visualize().

Optional extras

Install only what you need:

pip install "freshdata-cleaner[ml]"

Adds scikit-learn for KNN imputation and IsolationForest outlier detection (used in strategy="aggressive").

pip install "freshdata-cleaner[enterprise]"

Adds polars, pyarrow, requests, pyyaml for the enterprise layer: fuzzy clustering, PII masking, semantic validation, trust scoring, OpenLineage metadata, and the batch CLI.

pip install "freshdata-cleaner[all]"

All extras above plus cleanlab for ML label-noise detection.

pip install "freshdata-cleaner[polars]"

Pass a Polars DataFrame to fd.clean and get a Polars DataFrame back.

Python 3.9 on Linux aarch64: the privacy extra builds from source

freshdata-cleaner[privacy] (and [all]) pulls in spaCy through Presidio. On Python 3.9 spaCy is capped at 3.8.7, which requires thinc>=8.3.4,<8.4, and neither thinc 8.3.4 nor blis 1.2.0 publishes a cp39 Linux aarch64 wheel. pip therefore compiles thinc and blis from source there, which needs a C/C++ toolchain and takes several minutes. Python 3.10+ or x86-64 installs use prebuilt wheels.

Verify the installation

python -c "import freshdata as fd; print(fd.__version__)"

import pandas as pd
import freshdata as fd

df = pd.DataFrame({"a": [1, 2, 2, None], "b": [" x ", "y", "y", "z"]})
print(fd.clean(df))

Note on naming

The PyPI distribution is freshdata, but the import name is simply freshdata — so you install one and import the other:

pip install freshdata-cleaner

import freshdata as fd