JarCasting
JarCasting JarCasting

Cartogram geo

The jacart geo project for integrating geotools and jts with cartogram creation

License

License
Categories

Categories

Geo Business Logic Libraries Geospatial
GroupId

GroupId

de.dandit
ArtifactId

ArtifactId

cartogram-geo
Last Version

Last Version

0.2.0
Release Date

Release Date
Type

Type

jar
Description

Description

Cartogram geo
The jacart geo project for integrating geotools and jts with cartogram creation
Source Code Management

Source Code Management

https://github.com/DanDits/jacart/tree/master/cartogram-geo

Download cartogram-geo

Filename Size
cartogram-geo-0.2.0.pom
cartogram-geo-0.2.0.jar 18 KB
cartogram-geo-0.2.0-sources.jar 9 KB
cartogram-geo-0.2.0-javadoc.jar 413 KB
Browse

How to add to project

<!-- https://jarcasting.com/artifacts/de.dandit/cartogram-geo/ -->
<dependency>
    <groupId>de.dandit</groupId>
    <artifactId>cartogram-geo</artifactId>
    <version>0.2.0</version>
</dependency>
// https://jarcasting.com/artifacts/de.dandit/cartogram-geo/
implementation 'de.dandit:cartogram-geo:0.2.0'
// https://jarcasting.com/artifacts/de.dandit/cartogram-geo/
implementation ("de.dandit:cartogram-geo:0.2.0")
'de.dandit:cartogram-geo:jar:0.2.0'
<dependency org="de.dandit" name="cartogram-geo" rev="0.2.0">
  <artifact name="cartogram-geo" type="jar" />
</dependency>
@Grapes(
@Grab(group='de.dandit', module='cartogram-geo', version='0.2.0')
)
libraryDependencies += "de.dandit" % "cartogram-geo" % "0.2.0"
[de.dandit/cartogram-geo "0.2.0"]

Dependencies

compile (3)

Group / Artifact Type Version
de.dandit : cartogram-core jar 0.2.0
org.geotools » gt-main jar 22.3
org.geotools » gt-geojson jar 22.3

test (1)

Group / Artifact Type Version
org.junit.jupiter : junit-jupiter jar 5.6.0

Project Modules

There are no modules declared in this project.

JaCart

This project is based on the amazing project GoCart by Gastner, Seguy & More.

Gastner MT, Seguy V, More P. Fast flow-based algorithm for creating density-equalizing map projections. Proc Natl Acad Sci USA 115(10):E2156–E2164 (2018)

It allows you to transform a list of polygonal regions that are related to some positive numeric value as a cartogram. Those represent the related value by the regions' area, thus need to change the original geometry. We implement area contiguous cartograms that retain the original topology (i.e. the dual map of the original map and transformed map are isomorphic). Other cartogram variants like non contiguous Dorling cartograms may follow.

Examples

TODO provide some examples, link to blogpost

Project Structure

The project is in a beta phase and has a version 0.2.0 released to maven central under groupId "de.dandit" and is targeted to support Java 11. It is structured as follows:

  • cartogram-core: Basic cartogram logic and API. No external dependencies.
  • cartogram-geo: Offers utility methods for converting from/to geotools features, jts geometries and exporting results.

The cartogram-core API is based on the following: Create and supply a MapFeatureData and configure the execution using a CartogramConfig. For an example usage see the CartogramApiTest. Use the CartogramApi to create the desired cartogram. On success the CartogramResult contains information about the convergence, the transformed regions with their polygons and the used projection.

The cartogram-geo project does not have a clearly defined API yet and is more a utility and convenience library. This may change in future versions.

Performance and Comments

To give some ideas about convergence performance:

  • For usual input without extreme values and a 4 core machine using the commonPool parallelism configuration calculation takes around 2-3 seconds.
  • Using higher or lower resolution polygons does not significantly impact the execution time (if the point count does not exceed the chosen grid size which is 512x512 by default).
  • Scaling small regions to a very big size or vice versa can cause slow performance or even cause convergence to fail
  • The convergence is slower than the C based implementation GoCart, but not by a lot.
  • You can configure and tune the used FFT implementation, though the fourier transformations are not the bottleneck of execution time.

Open TODOs:

  • Created polygons can contain self intersections if original polygons were already contained bottlenecks, narrow areas, unlucky fractal coasts or line segments with a relatively long distance line segments. If you require the output to be valid geometries you should tune your input to introduce new coordinates within line segments that would be further apart than of the internally used grid cell size.
  • Identify bottlenecks and low hanging fruits: The bilinear interpolation needs to be improved or called less frequently.

Versions

Version
0.2.0
0.1.0