18 KiB

Raw Blame History

Creating a map

The topics that we will cover in this chapter include:

Creating a map
Define GeoJSON
Use a projection
Generate a <path> using a projection and the GeoJSON data
Create a rectangular overlay based on coordinates entered into a form
Draw a rectangle on the map and display its map coordinates
Draw a polygon on the map and display its map coordinates

In this section we'll generate <path> elements from GeoJSON data that will draw a map of the world

Define GeoJSON

GeoJSON is just JSON data that has specific properties that are assigned specific data types. Here's an example:

{
    "type": "Feature",
    "geometry": {
        "type": "Point",
        "coordinates": [125.6, 10.1]
    },
    "properties": {
        "name": "Dinagat Islands"
    }
}

In this example, we have one Feature who's geometry is a Point with the coordinates [125.6, 10.1]. It has "Dinagat Islands" as its name. Each Feature follows this general structure:

{
    "type": STRING,
    "geometry": {
        "type": STRING,
        "coordinates": ARRAY
    },
    "properties": OBJECT
}

We can also have a Feature Collection which is many Features grouped together in a features array:

{
    "type": "FeatureCollection",
    "features": [
        {
            "type": "Feature",
            "geometry": {
                "type": "Point",
                "coordinates": [102.0, 0.5]
            },
            "properties": {
                "prop0": "value0"
            }
        },
        {
            "type": "Feature",
            "geometry": {
                "type": "LineString",
                "coordinates": [
                    [102.0, 0.0], [103.0, 1.0], [104.0, 0.0], [105.0, 1.0]
                ]
            },
            "properties": {
                "prop0": "value0",
                "prop1": 0.0
            }
        },
        {
            "type": "Feature",
            "geometry": {
                "type": "Polygon",
                "coordinates": [
                    [
                        [100.0, 0.0], [101.0, 0.0], [101.0, 1.0],
                        [100.0, 1.0], [100.0, 0.0]
                    ]
                ]
            },
            "properties": {
                "prop0": "value0",
                "prop1": { "this": "that" }
            }
        }
    ]
}

This basically follows the form:

{
    "type": "FeatureCollection",
    "features": ARRAY
}

The features property is an array of feature objects which we've defined previously.

Set up the HTML

Let's set up a basic D3 page:

<!DOCTYPE html>
<html lang="en" dir="ltr">
<head>
    <meta charset="utf-8">
    <title></title>
    <script src="https://d3js.org/d3.v7.min.js" charset="utf-8"></script>
    <script src="https://cdn.rawgit.com/mahuntington/mapping-demo/master/map_data.js" charset="utf-8"></script>
</head>
<body>
    <svg></svg>
    <script src="app.js" charset="utf-8"></script>
</body>
</html>

The only thing different from the setup that we've used in previous chapters is this line:

<script src="https://cdn.rawgit.com/mahuntington/mapping-demo/master/map_data.js" charset="utf-8"></script>

This just loads an external javascript file which sets our GeoJSON data to a variable. Here's what the beginning of it looks like:

var map_json = {
    type: "FeatureCollection",
    features: [
        {
            type: "Feature",
            id: "AFG",
            properties: {
                name: "Afghanistan"
            },
            geometry: {
                type: "Polygon",
                coordinates: [
                    //lots of coordinates
                ]
            }
        }
        // lots of other countries
    ]
}

Note that the map_json variable is just a JavaScript object that adheres to the GeoJSON structure (it adds an id property which is optional). This is very important. If the object didn't adhere to the GeoJSON structure, D3 would not work as it should.

In production, you would probably make an AJAX call to get this data, or at the very least, create your own geoJSON file similar to the one being hosted on rawgit.com. The setup above was created to make learning easier by decreasing the complexity associated with AJAX.

Use a projection

Now let's start our app.js file:

const width = 960;
const height = 490;

d3.select('svg')
    .attr('width', width)
    .attr('height', height);

At the bottom of app.js let's add:

const worldProjection = d3.geoEquirectangular();

This generates a projection, which governs how we're going to display a round world on a flat screen. There's lots of different types of projections we can use: https://github.com/d3/d3-geo/blob/master/README.md#azimuthal-projections

The line above tells D3 to create an equirectangular projection (https://github.com/d3/d3-geo/blob/master/README.md#geoEquirectangular)

Generate a `<path>` using a projection and the GeoJSON data

Now that we have our projection, we're going to generate <path> elements for each data element in the map_json.features array. Then we set the fill of each element to #099. Add this at the end of app.js:

d3.select('svg').selectAll('path')
    .data(map_json.features)
    .enter()
    .append('path')
    .attr('fill', '#099');

Here's what it should look like at the moment if we open index.html in Chrome and view the elements tab in the developer tools:

We created the path elements, but they each need a d attribute which will determine how they're going to drawn (i.e. their shape).

We want something like:

d3.selectAll('path').attr('d', (datum, index)=>{
    //somehow use datum to generate the value for the 'd' attributes
});

Writing the kind of code described in the comment above would be very difficult. Luckily, D3 can generate that entire function for us. All we need to do is specify the projection that we created earlier. At the bottom of app.js add the following:

const dAttributeFunction = d3.geoPath()
    .projection(worldProjection);

d3.selectAll('path').attr('d', dAttributeFunction);

geoPath() generates the function that we'll use for the d attribute, and projection(worldProjection) tells it to use the worldProjection var created earlier so that the path elements appear as an equirectangular projection like this (This is helpful because we can use different projections to view a round world on a flat screen in different ways):

Create a rectangular overlay based on coordinates entered into a form

First, let's create a form for the user to enter lat/lng coordinates for the top/left and bottom/right portions of the box:

<form>
    Top Left:
    <input type="text" placeholder="latitude"/>
    <input type="text" placeholder="longitude"/>
    <br/>
    Bottom Right:
    <input type="text" placeholder="latitude"/>
    <input type="text" placeholder="longitude"/>

    <input type="submit"/>
</form>

Create a submit handler for the form
Prevent its default behavior of submitting so we stay on the page
Remove any rectangles that may have previously been drawn

d3.select('form').on('submit', (event) => { //set up a submit handler on the form
	event.preventDefault(); // stop the form from submitting
})

At the bottom of the submit handler, retrieve what the user entered in the inputs for the top left coordinate:

const lat1 = d3.select('input:first-child').property('value'); //get first latitude input
const lng1 = d3.select('input:nth-child(2)').property('value'); //get first longitude input

Now we'll use worldProjection() like our previous scales to convert a data point into pixel coordinates. It takes as a parameter, an array containing a latitude value and a longitude value ([lat,lng]). It returns an array containing x and y coordinates ([x,y]). Place the following at the bottom of your submit handler:

const location1 = worldProjection([lat1, lng1]); //convert these into pixel coordinates
const x1 = location1[0]; //reassign for ease of reading
const y1 = location1[1];

Do the same for the second for bottom right coordinate inputs (keep in mind, in the html the <br/> tag is the 3rd child of its parent, so the 3rd and 4th inputs are actually nth-child(4) and nth-child(5):

const lat2 = d3.select('input:nth-child(4)').property('value'); //get second latitude input (the <br/> is :nth-child(3)
const lng2 = d3.select('input:nth-child(5)').property('value'); //get second longitude input
const location2 = worldProjection([lat2, lng2]); //convert these into pixel coordinates
const x2 = location2[0]; //reassign for eas of reading
const y2 = location2[1];

Lastly, create the rectangle by appending it to the SVG, filling it with black, assigning the top left x/y coordinates and computing the height/width using the x2 and y2 coordinates:

d3.select('svg').append('rect') // append a rectangle to the svg
	.attr('fill', '#000') //make it black
	.attr('x', x1) //assign top left x location
	.attr('y', y1) //assign top left y location
	.attr('width', x2-x1) //compute width from bottom right x coordinate
	.attr('height', y2-y1) //compute height from bottom right y coordinate

The finished submit handler:

d3.select('form').on('submit', (event) => { //set up a submit handler on the form
	event.preventDefault(); // stop the form from submitting

	d3.selectAll('rect').remove(); //remove any previously drawn rectangles

	const lat1 = d3.select('input:first-child').property('value'); //get first latitude input
	const lng1 = d3.select('input:nth-child(2)').property('value'); //get first longitude input
	const location1 = worldProjection([lat1, lng1]); //convert these into pixel coordinates
	const x1 = location1[0]; //reassign for ease of reading
	const y1 = location1[1];

	const lat2 = d3.select('input:nth-child(4)').property('value'); //get second latitude input (the <br/> is :nth-child(3)
	const lng2 = d3.select('input:nth-child(5)').property('value'); //get second longitude input
	const location2 = worldProjection([lat2, lng2]); //convert these into pixel coordinates
	const x2 = location2[0]; //reassign for eas of reading
	const y2 = location2[1];

	d3.select('svg').append('rect') // append a rectangle to the svg
		.attr('fill', '#000') //make it black
		.attr('x', x1) //assign top left x location
		.attr('y', y1) //assign top left y location
		.attr('width', x2-x1) //compute width from bottom right x coordinate
		.attr('height', y2-y1) //compute height from bottom right y coordinate
});

Draw a rectangle on the map and display its map coordinates

First create a table in our HTML. Once the user draws the rectangle on the map, we'll have the app fill the data in here appropriately:

<table border="1">
	<tr>
		<th>lat</th>
		<th>lng</th>
	</tr>
	<tr>
		<td></td>
		<td></td>
	</tr>
	<tr>
		<td></td>
		<td></td>
	</tr>
</table>

Now set up a drag behavior and attach it to the SVG:

const dragBehavior = d3.drag() //set up the behavior
	.on('start', (event) => { //create the start handler
	})
	.on('drag', (event) => { //create the drag handler
	})

d3.select('svg').call(dragBehavior); //atach dragBehavior to the svg

Next, let's use worldProjection.invert() to take visual x/y coordinates and convert them into lat/lng coordinates. Has parameters, it takes an array containing x and y pixel coordinates [x,y], and it returns and array containing lat and lang coordinates [lat,lng]. We'll use worldProjection.invert([event.x, event.y])[0] to get the latitude and worldProjection.invert([event.x, event.y])[1] to get the longitude. We'll then insert those values into the cells in tr:nth-child(2) (nth-child(1) is the header tr). At the bottom of the start handler, add the following:

d3.select('table tr:nth-child(2) td:first-child') //select the top left cell of the table. The table header is tr:nth-child(1)
	//use worldProject to convert the x/y pixel coordinates of the mouse click into lat/lng map coordinates
	//this takes as params [x,y] and returns [lat,lng]
	//insert the first element of the map coords array (lat) into the html of the table cell
	.html(worldProjection.invert([event.x, event.y])[0]) 
d3.select('table tr:nth-child(2) td:last-child') //select the top right cell of the table
	//insert the second element of the map coords array (lng) into the html of the table cell
	.html(worldProjection.invert([event.x, event.y])[1])

Now we want to begin drawing the rectangle. We'll start with a rectangle of 0 height and 0 width, at the point the user clicked. Add the following at the bottom of the start hander:

d3.selectAll('rect').remove(); //remove any previously drawn rectangles
d3.select('svg').append('rect') //draw a rectangle of 0 height/width at the point where the drag started
	.attr('x', event.x)
	.attr('y', event.y);

When the user drags on the SVG, we want to update the rectangle's height/width and add lat/lng coordinates for the bottom right of the square in the table. First, let's take care of the table. Add the following to the drag handler (this is almost identical to the start handler, except we use tr:nth-child(3)):

d3.select('table tr:nth-child(3) td:first-child') //select the bottom left cell of the table
	//insert the first element of the map coords array (lat) into the html of the table cell
	.html(worldProjection.invert([event.x, event.y])[0])
d3.select('table tr:nth-child(3) td:last-child') //select the bottom right cell of the table
	//insert the second element of the map coords array (lng) into the html of the table cell
	.html(worldProjection.invert([event.x, event.y])[1])

To update the rectangle's width/height, we need to calculate it based on the current event's x/y position and the rectangle's starting x/y position. Please add the following at the bottom of the drag handler:

//computer width of the box based on current x value and current rect's x value
d3.select('rect').attr('width', event.x-d3.select('rect').attr('x'))
//computer height of the box based on current y value and current rect's y value
d3.select('rect').attr('height', event.y-d3.select('rect').attr('y'))

The finished drag behavior:

const dragBehavior = d3.drag() //set up the behavior
	.on('start', (event) => { //create the start handler
		d3.select('table tr:nth-child(2) td:first-child') //select the top left cell of the table. The table header is tr:nth-child(1)
			//use worldProject to convert the x/y pixel coordinates of the mouse click into lat/lng map coordinates
			//this takes as params [x,y] and returns [lat,lng]
			//insert the first element of the map coords array (lat) into the html of the table cell
			.html(worldProjection.invert([event.x, event.y])[0]) 
		d3.select('table tr:nth-child(2) td:last-child') //select the top right cell of the table
			//insert the second element of the map coords array (lng) into the html of the table cell
			.html(worldProjection.invert([event.x, event.y])[1])

		d3.selectAll('rect').remove(); //remove any previously drawn rectangles
		d3.select('svg').append('rect') //draw a rectangle of 0 height/width at the point where the drag started
			.attr('x', event.x)
			.attr('y', event.y);
	})
	.on('drag', (event) => { //create the drag handler
		d3.select('table tr:nth-child(3) td:first-child') //select the bottom left cell of the table
			//insert the first element of the map coords array (lat) into the html of the table cell
			.html(worldProjection.invert([event.x, event.y])[0])
		d3.select('table tr:nth-child(3) td:last-child') //select the bottom right cell of the table
			//insert the second element of the map coords array (lng) into the html of the table cell
			.html(worldProjection.invert([event.x, event.y])[1])

		//computer width of the box based on current x value and current rect's x value
		d3.select('rect').attr('width', event.x-d3.select('rect').attr('x'))
		//computer height of the box based on current y value and current rect's y value
		d3.select('rect').attr('height', event.y-d3.select('rect').attr('y'))
	})

d3.select('svg').call(dragBehavior); //atach dragBehavior to the svg

Draw a polygon on the map and display its map coordinates

<!DOCTYPE html>
<html lang="en" dir="ltr">
	<head>
		<meta charset="utf-8">
		<title></title>
		<script src="https://d3js.org/d3.v7.min.js" charset="utf-8"></script>
		<script src="https://cdn.rawgit.com/mahuntington/mapping-demo/master/map_data.js" charset="utf-8"></script>
	</head>
<body>
	<svg></svg>
	<table>
		<thead>
			<tr>
				<th>lat</th>
				<th>lng</th>
			</tr>
		</thead>
		<tbody></tbody>
	</table>
	<script src="app.js" charset="utf-8"></script>
</body>
</html>

const width = 960;
const height = 490;

d3.select('svg')
	.attr('width', width)
	.attr('height', height);

const worldProjection = d3.geoEquirectangular();
const dAttributeFunction = d3.geoPath()
	.projection(worldProjection);

d3.select('svg').selectAll('path')
	.data(map_json.features)
	.enter()
	.append('path')
	.attr('fill', '#099')
	.attr('d', dAttributeFunction);

let points = [];
d3.select('svg').on('click', (event)=>{
	points.push([event.x, event.y])
	const path = d3.path();
	path.moveTo(points[0][0], points[0][1])
	for(let i = 1; i < points.length; i++){
		path.lineTo(points[i][0], points[i][1])
	}
	path.closePath()
	d3.select('#polygon').remove()
	d3.select('svg')
		.append('path')
		.attr('id', 'polygon')
		.attr('fill', 'black')
		.attr('d', path)
})
d3.select('body').on('keydown', (event)=>{
	d3.select('tbody').html('');
	const coords = points.map(point => worldProjection.invert(point))
	for(coord of coords){
		const row = d3.select('tbody')
			.append('tr')
		row.append('td')
			.html(coord[0])
		row.append('td')
			.html(coord[1])
	}
	points = [];
})

Conclusion

In this section we've covered how to use D3 to create a projection and render GeoJSON data as a map, and we've learned about using different projects to visualize the world. This can be helpful when displaying populations or perhaps average rainfall of various regions. Congratulations! You've made it to the end of this book. Now go off and create amazing visualizations.

18 KiB Raw Blame History