Next-Gen Styling for Data-Driven Maps, AKA CartoCSS preprocessor.
tl;dr Enables adding functions to CartoCSS that can be evaluated asynchronously.
It's all about ramps.
You can create color and symbol size ramps with just a single line of code. You no longer need to worry about calculating the bins for your thematic map, if your data changes your CartoCSS changes.
In CartoCSS you usually assign values to properties and apply filters in order to change those values based on some data attributes.
The general form for properties and filters is:
#selector {
property: value;
[filter] {
property: value;
}
}
An example of a filter based on price attribute:
#selector {
marker-width: 10;
[price > 100] {
marker-width: 20;
}
}
Turbo-Carto high-level API for ramps is as follows:
#selector {
property: ramp([attribute], (...values), (...filters), mapping);
}
Where:
property
is the CartoCSS property you want to create.[attribute]
usually is a column/key name from your dataset.(...values)
is whatproperty
is gonna get for different filters.(...filters)
is howproperty
is gonna get the different values.mapping
is the type of filter that will be applied: <, <=, >, >=, =.
So for the previous example you could write (see examples/readme/example-0.css):
#selector {
marker-width: ramp([price], (10, 20), (100));
}
In this case, the first value is the default value.
If you want to have a category map, to generate a CartoCSS like:
#selector {
marker-fill: red;
[room_type = "Private Room"] {
marker-fill: green;
}
}
You can use the same approach but specifying the mapping type to be an equality (see examples/readme/example-1.css):
#selector {
marker-fill: ramp([room_type], (green, red), ("Private room"), =);
}
See that in this case the last value is the default value, and if the number of values is equal to the number of filters it won't have a default value, like in (see examples/readme/example-2.css):
#selector {
marker-fill: ramp([room_type], (green, red), ("Private room"), =);
}
That's nice, but it is still unlinked from the actual data/attributes.
<
and<=
: Last provided value will be the default value.=
: Last provided value will be the default value, if the number of values is equal to the number of filters it won't have a default value.>
and>=
: First provided value will be the default value.
To generate ramps based on actual data you have to say what kind of quantification you want to use, for that purpose Turbo-Carto provides some shorthand methods that will delegate the filters computation to different collaborators.
Let's say you want to compute the ramp using jenks as quantification function (see examples/readme/example-3.css):
#selector {
marker-width: ramp([price], (10, 20, 30), jenks());
}
Or generate a category map depending on the room_type property (see examples/readme/example-4.css):
#selector {
marker-fill: ramp([room_type], (red, green, blue), category(2));
}
You can override the mapping if you know your data requires an more strict filter (see examples/readme/example-5.css):
#selector {
marker-width: ramp([price], (10, 20, 30), jenks(), >=);
}
Shorthand methods include:
category()
: default mapping is=
.equal()
: default mapping is>
.headtails()
: default mapping is<
.jenks()
: default mapping is>
.quantiles()
: default mapping is>
.
For color schemes, there are a couple of handy functions to retrieve color palettes: cartocolor
and colorbrewer
.
You can create a choropleth map using Reds color palette from colorbrewer (see examples/readme/example-6.css):
#selector {
polygon-fill: ramp([avg_price], colorbrewer(Reds), jenks());
}
Or a category map using Bold color palette from cartocolor (see examples/readme/example-7.css):
#selector {
polygon-fill: ramp([room_type], cartocolor(Bold), category(4));
}
Go to CARTOcolors website to discover the wide diversity of color schemes that are available.
Sometimes is really useful to have a continuous range that can be split in as many bins as desired, so you don't have to worry about how many values you have to provide for your calculated filters. You only specify the start and the end, and the values are computed linearly (see examples/readme/example-8.css):
#selector {
marker-width: ramp([price], range(4, 40), equal(5));
}
It is kind of similar to color palettes where depending on your number of filters you get the correct number of values.
- Node >=0.10
- npm >=2.x
To install the rest of the dependencies just run:
npm install
Current examples expect to have CartoDB's SQL API running at http://development.localhost.lan:8080/. It should have a
publicly accessible populated_places_simple
table.
Examples using a cli tool:
$ ./tools/cli.js examples/populated_places.css # will use a dummy datasource, check tools/cli.js source
$ ./tools/cli.js examples/populated_places.css --datasource sql --query 'select * from populated_places_simple'
Visit examples/index.html
, to test different styles using CartoDB's SQL API as datasource.
Tests suite can be run with:
npm test
That will check code style and run the tests.