Semantic Link Functions for Geopandas

0.14.0 · active · verified Wed Apr 15

Semantic link functions for Geopandas enables conversion of a FabricDataFrame to a GeoDataFrame. This package is part of the broader Microsoft Fabric Semantic Link ecosystem, which facilitates connecting Power BI datasets with Synapse Data Science. It is actively maintained, with new versions often released in conjunction with other `semantic-link` packages.

Warnings

breaking Installing `semantic-link` (and thus implicitly `semantic-link-functions-geopandas`) with `pandas >= 3.0` can cause runtime errors due to breaking changes in Pandas' internal API. Microsoft is actively working on a fix.
Fix: Downgrade `pandas` to a version less than 3.0 (e.g., `pip install 'pandas<3.0'`) or ensure you are using a patched `semantic-link` version if available. Alternatively, install `semantic-link-sempy` directly instead of the `semantic-link` meta-package.
gotcha The full Semantic Link feature set, including the semantic functions and `FabricDataFrame`'s capabilities, is primarily designed and supported for use within Microsoft Fabric notebooks and environments. While the Python packages can be installed locally, their full integration with Power BI semantic models is specific to the Fabric platform.
Fix: For full functionality and integration, use this library within a Microsoft Fabric workspace. Local usage will lack direct connectivity to Power BI semantic models.
deprecated The underlying `geopandas` library has enforced several deprecations in recent major versions (e.g., `geopandas.datasets` module removed, `geopandas.io.read_file` deprecated in favor of `geopandas.read_file`). Users migrating or updating `geopandas` versions independently of `semantic-link-functions-geopandas` should be aware of these changes.
Fix: Refer to the `geopandas` changelog for specific deprecation details and migration paths. Update your `geopandas`-specific code to use the modern API calls.

Install

pip install semantic-link-functions-geopandas Install specific package
pip install semantic-link Install meta-package (includes geopandas functions)

Imports

FabricDataFrame

from sempy.fabric import FabricDataFrame

Quickstart

Creates a sample FabricDataFrame with latitude and longitude, then converts it into a GeoDataFrame using the `to_geopandas` semantic function. This demonstrates the core functionality of mapping semantic columns to a geospatial object.

import pandas as pd
from sempy.fabric import FabricDataFrame

# Example FabricDataFrame (usually obtained from a Power BI dataset in Fabric)
df = FabricDataFrame(
    {
        "country": ["US", "AT"],
        "lat": [40.7128, 47.8095],
        "long": [-74.0060, 13.0550]
    },
    column_metadata={
        "lat": {"data_category": "Latitude"},
        "long": {"data_category": "Longitude"}
    },
    pd_dataframe=pd.DataFrame()
)

# Convert to GeoDataFrame
df_geo = df.to_geopandas(lat_col="lat", long_col="long")

print(df_geo.head())
print(df_geo.geometry.name)

view raw JSON →