UDL (UrbanDataLayer) is a suite of standard data structure and pipeline for city data engineering, which processes city data from various raw data into a unified data format.

UDL is featured for:

Unified standardized formats for city data: five data layers (grid, graph, point, polygon and linestring).
User-friendly APIs of data processing: scheme transformation, granularity alignment and feature fusion.

Installation¶

It is recommended to use pip for installation. Please make sure the latest version is installed, as UDL may be updated frequently:

pip install udlayer          # pip install
pip install --upgrade udlayer  # update if needed

Note

If some error about gdal appears when installing, please install geopandas package manually (from wheel or use conda). It may be caused by the dependency of geopandas. Please ensure that the latest udlayer version is installed.

Note on PyG and DGL Installation: Some conversion methods (in utils) depends on torch, torch_geometric and DGL. To streamline the installation, UDL does NOT install these libraries for you. Please install them from the above links if you need to convert the graph data into these two formats, otherwise you can ignore this note.

Quick Start¶

We provide complete sample process of “Identification of administrative boundaries” and “pm2.5 prediction” cases to help you get started with UDL in the UDL repository.