COASTGON DATASET
Regional coastal characteristics of the Earths coastline between -37 and 
48 degrees latitude on a hexagonal cell grid.

DATE OF RELEASE
28 December 2023

CONTACT INFORMATION
Jakob C. Christiaanse
Delft University of Technology, Department of Hydraulic Engineering
J.C.Christiaanse@tudelft.nl

OVERVIEW
This dataset contains the global coastgon grid, covering the Earth's coastline
between -37 and 48 degrees latitude in a hexagonal cell grid. Each Coastgon 
(hexagonal coastal cell) has a diameter of around 50 km, although this varies 
(with almost all coastgon diameters between 45 and 55 km). Each coastgon is 
characterized by 22 coastal indicators derived from state-of-the-art global 
datasets, spanning hydrodynamic, atmospheric, geophysical, habitat, and human 
processes. For more details on the derivation and background of the data, the 
reader is refered to the accompanying publication (Christiaanse et al., 2023).

REFERENCES
Christiaanse, J.C., Antolinez, J.A.A., Luijendijk, A.P., Athanasiou, P., Duarte, 
C.M., and Aarninkhof, S. (2024). Distribution of global sea turtle nesting 
explained from regional-scale coastal characteristics. Scientific Reports, 14(752). 
https://doi.org/10.1038/s41598-023-50239-5

CONTENTS
- coastgons_h3r4_meta.json:
   Contains the metadata belonging to the Coastgon dataset (information on variables)
- coastgons_h3r4.parquet: 
   Contains the coastgon dataset in a geoparquet file
- Coastgons_h3r4.csv: 
   Contains the Coastgon dataset in a CSV file

DATA FORMAT
Tabular data, provided in two formats:
- *.parquet file
- *.csv file

DATA STRUCTURE
The data set contains a *.parquet file (geoparquet) and a *.csv file, both contain 
the same data tabular geospatial data, where each row represents one coastgon, 
indexed by the unique H3 ID. The first three columns provide general information 
on each coastgon: the number of coastline transects it represents (n_gcc), and the 
two geometry objects that represent the geolocation of the coastgon itself (polygon)
and the coastline centroid (clc). Then follow the 22 columns which represent the 
coastal indicators for each coastgon, as explained in Christiaanse et al. 2023.
Please refer to the metadata file for more information on each column of the dataset.
The advantage of geoparquet is that it is faster and when read with GeoPandas in 
Python (see below), directly returns a GeoDataFrame, with Geometry column.

USAGE
Any software or programming language that can read *.parquet and/or *.csv files can 
be used to analyse the data (Python, R, C++, Julia). We have used Python (GeoPandas) 
for all processing and operations on the data. Required python packages are GeoPandas 
and Fiona (and ideally Shapely). The *.parquet file can then simply be read using 
GeoPandas.read_parquet(), which will return a GeoDataFrame with the data. For more
information on geoparquet, libraries to read the file, and advantages of geoparquet
see the website (https://geoparquet.org).

LICENSE
CC BY 4.0