D3 Force simulation in a grid

Question

I was wondering how it would be possible to modify Mike Bostock's example of a multi-force layout in order to try and get the force layout to group nodes in a grid.

So let us imagine that we have the following csv:

Name, Category1, Category2
1,1,1
2,1,2
3,1,1
4,2,2
5,3,1
6,1,4
7,5,5
8,1,5
9,2,4
10,3,3
11,4,4
12,4,5
13,3,4
14,1,2
15,1,1
16,2,2
17,3,1
18,2,1
19,4,5
20,3,1

For his kind of data I would like to have all the possible values of Category 1 as columns and all the possible values of Category 2 as rows and would like my nodes to automatically group in the "proper" cell depending on their values for Category 1 and Category 2.

I am just getting started with D3 and don't really know where to start. The example I pointed to is useful, but it's hard to know what to modify as the code has close to no comments.

Any help would be appreciated.

Answer 1

Forget that example: it uses D3 v3, which makes positioning the nodes way more complicated.

In D3 v4/v5 there are two convenient methods, forceX and forceY.

All you need to do is creating your scales, for instance using a point scale (the best choice here in my opinion):

var columnScale = d3.scalePoint()
  .domain(["1", "2", "3", "4", "5"])
  .range([min, max]);

var rowScale = d3.scalePoint()
  .domain(["1", "2", "3", "4", "5"])
  .range([min, max]);

And then use those scales in the simulation:

var simulation = d3.forceSimulation(data)
  .force("x", d3.forceX(function(d) {
    return columnScale(d.Category1)
  }))
  .force("y", d3.forceY(function(d) {
    return rowScale(d.Category2)
  }))

Here is a basic demo with the data you shared (I'm using a colour scale to highlight the different positions on the grid):

var csv = `Name,Category1,Category2
1,1,1
2,1,2
3,1,1
4,2,2
5,3,1
6,1,4
7,5,5
8,1,5
9,2,4
10,3,3
11,4,4
12,4,5
13,3,4
14,1,2
15,1,1
16,2,2
17,3,1
18,2,1
19,4,5
20,3,1`;

var data = d3.csvParse(csv);

var w = 250,
  h = 250;

var svg = d3.select("body")
  .append("svg")
  .attr("width", w)
  .attr("height", h);

var color = d3.scaleOrdinal(d3.schemeCategory10);

var columnScale = d3.scalePoint()
  .domain(dataRange(data, 'Category1')) // or ["1", "2", "3", "4", "5"]
  .range([30, w - 10])
  .padding(0.5);

var rowScale = d3.scalePoint()
  .domain(dataRange(data, 'Category2')) // or ["1", "2", "3", "4", "5"]
  .range([30, h - 10])
  .padding(0.5);

var simulation = d3.forceSimulation(data)
  .force("x", d3.forceX(function(d) {
    return columnScale(d.Category1)
  }))
  .force("y", d3.forceY(function(d) {
    return rowScale(d.Category2)
  }))
  .force("collide", d3.forceCollide(6))

var nodes = svg.selectAll(null)
  .data(data)
  .enter()
  .append("circle")
  .attr("r", 5)
  .attr("fill", function(d) {
    return color(d.Category1 + d.Category2)
  });

var xAxis = d3.axisTop(columnScale)(svg.append("g").attr("transform", "translate(0,30)"));

var yAxis = d3.axisLeft(rowScale)(svg.append("g").attr("transform", "translate(30,0)"));

simulation.on("tick", function() {
  nodes.attr("cx", function(d) {
      return d.x
    })
    .attr("cy", function(d) {
      return d.y
    })
});

function dataRange(records, field) {
  var min = d3.min(records.map(record => parseInt(record[field], 10)));
  var max = d3.max(records.map(record => parseInt(record[field], 10)));
  return d3.range(min, max + 1);
};

svg {
  background-color: floralwhite;
  border: 1px solid gray;
}

<script src="https://d3js.org/d3.v5.min.js"></script>

PS: In both scales I'm using strings in the domain because d3.csv will load your data as strings, not numbers. Change that according to your needs.