blackmarble-ntl-toolkit 0.1.2

A Python toolkit for retrieving and preprocessing NASA's Black Marble Nighttime Light (NTL) data using Google Earth Engine, xarray, and Dask.

Black Marble NTL Toolkit

Black Marble NTL Toolkit is a Python package designed to streamline the retrieval, cleaning, aggregation, and visualization of NASA's Nighttime Light (NTL) data.

Deploying NTL data for conflict monitoring or socioeconomic research demands a combination of deep domain expertise and advanced technical proficiency to retrieve, process, and analyze large-scale datasets. Consequently, effective research relies on close collaboration between conflict researchers, machine learning specialists, and data engineers. To lower these technical barriers and facilitate independent experimentation, the codebase contains helper functions designed to streamline the retrieval, cleaning, and visualization of NTL data.

📚 Documentation

Full documentation, tutorials, and API reference can be found here:
🔗 https://jankokla.github.io/blackmarble-ntl-toolkit/

---

✨ Main Functionality

1. Earth Engine Authentication & Data Retrieval

Before retrieving data, you must authenticate and initialize the Google Earth Engine Python API. Once initialized, you can seamlessly pull down daily and annual NASA Black Marble products.

from blackmarble_toolkit.utils import initialize_ee
from blackmarble_toolkit.retrieval import BlackMarbleRetriever

initialize_ee(project_name="your-gee-project-id")

retriever = BlackMarbleRetriever()

raw_ds = retriever.get_data(
    product="VNP46A2",
    start_date="2022-01-01", 
    end_date="2022-12-31",
    region=geometry
)

2. Preprocessing Pipeline

Create a robust, memory-efficient pipeline leveraging Dask and Xarray to filter out clouds, snow, and background noise, and correct geometric or angular distortions.

from blackmarble_toolkit.pipeline import NTLPipeline
from blackmarble_toolkit.methods import filters, angular, geometric, imputation

steps = [
    filters.CloudSnowFilter(),
    angular.QuadraticVZACorrection(),
    geometric.AveragePooling2D(filter_size=(3, 3)),
    imputation.LinearInterpolationGapFilling()
]

pipeline = NTLPipeline(steps)
processed_ds = pipeline.run(raw_ds, cache_intermediates=True)

3. Spatial Aggregation

Map your pixel-level NTL radiance to custom geometries (like administrative regions) for time-series analysis.

import geopandas as gpd

regions = gpd.read_file("regions.geojson")
regions = regions.reset_index(names="region_id")

aggregated_results = pipeline.aggregate(
    gdf=regions,
    geo_id_col="region_id"
)

4. Exporting to CSV

If you prefer doing downstream analysis and plotting in pandas or R, you can easily export the aggregated results to a CSV file.

df = pipeline.to_csv("ntl_aggregated_results.csv")

🛠️ Installation

You can install the toolkit directly via PyPI:

pip install blackmarble-ntl-toolkit

For more details on setting up your environment and configuring Earth Engine, please consult the Installation Guide.