Urbanomy is a research toolkit from the Institute of Design and Urban Studies (IDU) for assessing land parcels from spatial, economic, and scenario perspectives. The library bundles geospatial feature engineering, CatBoost-based land-price modelling, multi-criteria investment scoring, and socio-economic ripple analysis into one reproducible codebase.
- Land-value modelling – build spatial lag features, estimate land prices, and visualise impacts with pre-trained CatBoost models (urbanomy.methods.land_value_modeling).
- Scenario planning – modify individual blocks, recompute indicators, and plot price deltas for what-if cases (urbanomy.methods.land_value_modeling.scenario_modification).
- Investment attractiveness – aggregate cashflows, NPV/IRR/ROI metrics, and weighted spatial potential into a combined INV score (urbanomy.methods.investment_potential).
- Socio-economic footprint – quantify fiscal and employment effects during construction and operation phases with SEREstimator.
- Robust data validation – Pandera-based schemas for tabular and GeoDataFrame inputs (urbanomy.utils).
Urbanomy requires Python 3.10 and system libraries for GeoPandas (GEOS, GDAL, PROJ). We recommend working inside a virtual environment.
git clone https://github.com/iduclub/Urbanomy.git
cd Urbanomy
pip install -e .Optional dependency groups:
pip install -e '.[dev]'– formatting, linting, pre-commit hooks.pip install -e '.[test]'– pytest and coverage tooling.pip install -e '.[docs]'– Sphinx stack for building the documentation.
import pandas as pd
import geopandas as gpd
from catboost import CatBoostRegressor
from urbanomy.methods.land_value_modeling import (
LandDataPreparator,
LandPriceEstimator,
plot_land_price_maps,
)
scenario = gpd.read_file("data/scenario_blocks.geojson")
context = gpd.read_file("data/context_blocks.geojson")
accessibility = pd.read_pickle("data/accessibility.pkl")
preparator = LandDataPreparator(
scenario_blocks_source=scenario,
context_blocks_source=context,
accessibility_matrix_source=accessibility,
)
prepared_blocks = preparator.prepare()
model = CatBoostRegressor()
model.load_model("models/land_price.cbm")
estimator = LandPriceEstimator(model=model, blocks=prepared_blocks)
price_map = estimator.predict()
plot_land_price_maps(blocks_pred=price_map, scenario_blocks=scenario)Training the land-price model and generating spatial lag features
Price change map after applying the scenario
Comparing investments and prices for scenario and context blocks
from urbanomy.methods.investment_potential import (
InvestmentAttractivenessAnalyzer,
LandUseScoreAnalyzer,
prepare_investment_input,
)
# 1. Spatial potential scores per land-use
land_use_scores = LandUseScoreAnalyzer().compute_scores_long(prepared_blocks)
# 2. Merge scenario blocks with baseline scores
investment_ready = prepare_investment_input(
prepared_blocks,
base_gdf=land_use_scores,
scenario_flag_column="is_scn",
)
# 3. Calculate economic metrics and final INV index
BENCHMARKS = {
"residential": {"density": 1.2, "capex": 85_000, "rent_share": 0.75},
"business": {"density": 3.0, "capex": 120_000, "rent_share": 0.65},
# ... add other land-use profiles
}
analyzer = InvestmentAttractivenessAnalyzer(benchmarks=BENCHMARKS)
enriched_blocks, project_summary = analyzer.calculate_investment_metrics(investment_ready)Investment metrics dashboard with the final INV index
import pandas as gpd
from urbanomy.methods.ser import SEREstimator
ser = SEREstimator(
{
"population": 52_000,
"employment_base": 21_000,
"avg_wage_base": 70_000,
}
)
project_blocks = gpd.read_file('./data/test/blocks_investment.geojson')
ser_table = ser.compute(project_blocks)
print(ser_table)Reproducible pipelines live in examples/:
- examples/land_value_modeling/1_pricing_model_training.ipynb – model training workflow.
- examples/land_value_modeling/2_land_data_preparation.ipynb – block feature engineering.
- examples/land_value_modeling/3_land_price_modeling.ipynb – scenario pricing walkthrough.
- examples/investment_metrics.ipynb – investment attractiveness dashboard.
- examples/socio_economic_indicators.ipynb – socio-economic reporting.
The latest documentation is published automatically from main:
https://iduclub.github.io/Urbanomy/
Build the docs locally with:
pip install -e '.[docs]'
sphinx-build docs/source docs/buildWe use the supplied Makefile shortcuts:
make install-dev # install editable package with dev tools
make lint # pylint on src/urbanomy
make test # pytest tests/
make install-docs # editable install with docs extrasBefore opening a pull request:
- Format imports with
isortand code withblack(make format). - Ensure
make testpasses. - Run
sphinx-buildif documentation is affected.
Urbanomy is released under the BSD-3-Clause license. See LICENSE for the full text.
Maintainer: Maksim Natykin (mvin@itmo.ru)
Issues & feature requests: https://github.com/iduclub/Urbanomy/issues