Dana Silver @DanaRSilver GitHub

February 11, 2014

Compare Historically Popular names

Loading the data...
Share this chart:

Making the Chart


Margins, width, and height are standard. Years are on the x-axis, so a d3.time.scale is appropriate. Similarly, a linear scale is appropriate for the number of names on the y-axis. The category10 ordinal scale will assign a color to each line.

The axes need minimal setup. Each is scaled based on its respective x or y scale. Each line is created by assigning its x values to the years of the individual data points and the y values to the number of names (count) of the individual data points. Using a cardinal interpolation removes sharp tangents on the lines using a cardinal spline.

Finally, the svg element is given its height, width, and a inner group element to contain the actual chart.

var margin = {top: 30, right: 0, bottom: 30, left: 50},
    width = 840 - margin.left - margin.right,
    height = 500 - margin.top - margin.bottom;

var x = d3.time.scale()
    .range([0, width]);

var y = d3.scale.linear()
    .range([height, 0]);

var color = d3.scale.category10();

var xAxis = d3.svg.axis()

var yAxis = d3.svg.axis()

var line = d3.svg.line()
    .x(function(d) { return x(d.year) })
    .y(function(d) { return y(d.count) });

var svg = d3.select("#names")
    .attr("width", width + margin.left + margin.right)
    .attr("height", height + margin.top + margin.bottom)
    .attr("transform", "translate(" + margin.left + "," + margin.top + ")");


The data is hosted at files.danasilver.org/names1880-2012.csv. I obtained it from my Computer Science Information Visualization course. I request the data with d3.csv, use an accessor method to format each data point, and use a callback to manipulate the data.

Saving all the loaded data to the variable allData makes it easy to make other charts based on user input without requesting the data each time. The first time the data is loaded, I check if there is a names URL parameter from which to create the initial chart. If names is empty, I use the placeholder "Jim, Pam, Stanley" to create the chart.

The .on("progress") listens for d3.csv's progress event and updates the progress bar.

var allData;
d3.csv("//files.danasilver.org/names1880-2012.csv", function(d) {
  return {
    year: new Date(+d.Year, 0, 1),
    count: +d.Count,
    name: d.Name + ", " + d.Gender,
}, function(data) {
  allData = data;
  document.querySelector("#loading").innerHTML = "";

  paramNames = getParameterByName("names");
  if (paramNames === "") {
    makeChart("Jim, Pam, Stanley");
  } else {

}).on("progress", function(e) {
  if (d3.event.lengthComputable) {
    percent = (d3.event.loaded / d3.event.totalSize) * 100;
    document.querySelector("progress").value = percent;


For this application, it's useful to keep the chart logic in a separate function so it can be called each time the user enters some names. This function takes a list of comma separated names as its only parameter. It makes sense to annotate the source for this complicated and integral function rather than provide a high level overview. The commented code is below.

function makeChart(names) {

  // Crudely empty the contents of the chart
  document.querySelector("#names g").innerHTML = "";

  // Create an array from the string of comma separated names
  // One array is created with ', m' postpended to each name, the other
  // with ', f' to chart male and female names.
  // The male and female arrays are then concatenated.
  var namesArray = names.split(",").map(function(n) { return n.trim().toLowerCase(); }),
      namesArrayM = namesArray.map(function(n) { return n + ", m"}),
      namesArrayF = namesArray.map(function(n) { return n + ", f"}),
      namesSearchArray = namesArrayM.concat(namesArrayF);

  // Replace the share url with the new string of names
  var baseURL = window.location.protocol + "//" + window.location.host + window.location.pathname,
      newUrl = baseURL + "?names=" + namesArray.join(",");
  document.getElementById("chartUrl").value = newUrl;

  // Filter all the raw data using the names in name search array
  data = allData.filter(function(key) { return (namesSearchArray.indexOf(key.name.toLowerCase()) > -1) });

  // Nest the data based on name
  // This works similarly to SQL's GROUP BY, grouping all the elements with
  // the same key and listing the values for each distinct key.
  // More on nesting: https://github.com/mbostock/d3/wiki/Arrays#wiki-d3_nest
  data = d3.nest()
            .key(function(d) { return d.name; })

  // Sort the data by descending number of names (count)
  // The comparison keys (a and b) are names with associated data.
  // Sum all the counts for each name to obtain the total count for that name.
  // Though it has no bearing on the placement of the lines, this ensures the
  // names in the legend appear in order of decreasing total count.
  data.sort(function(a,b) {
    return d3.sum(b.values, function(d) { return d.count }) - d3.sum(a.values, function(d) { return d.count });

  // Assign the domain of the colors ordinal scale to the 'name' key of each data group
  // Now each name will map to a color
  color.domain(d3.keys(data[0]).filter(function(key) { return key === "name"; }));

  // Assign the x domain using the minimum and maximum years for each line
  // This means the chart may not show all years, but will probably be easier to read
    d3.min(data, function(d) { return d3.min(d.values, function(d) { return d.year }); }),
    d3.max(data, function(d) { return d3.max(d.values, function(d) { return d.year }); })

  // Assign the y domain from 0 to the maximum count value of all the data points
  y.domain([0, d3.max(data, function(d) { return d3.max(d.values, function(d) { return d.count }); })]);

  // Create the visual x-axis at the bottom of the chart
      .attr("class", "x axis")
      .attr("transform", "translate(0," + height + ")")

  // Create the visual y-axis at the bottom of the chart and append a label
      .attr("class", "y axis")
      .attr("transform", "rotate(-90)")
      .attr("y", 6)
      .attr("dy", ".71em")
      .style("text-anchor", "end")
      .text("# Names");

  // Create g element for each dataset
  // Each of these elements will contain a line
  var name = svg.selectAll(".name")
      .data(data, function(d) { return d.key; })
      .attr("class", "name");

  // Create the lines
  // The d attribute makes the actual line shape,
  // which comes from the line defined in the preparation
  // and is passed in the values from a particular line's dataset
      .attr("class", "line")
      .attr("d", function(d) { return line(d.values) })
      .style("stroke", function(d) { return color(d.key) })

  // Append a g element for the legend
  var legend = svg.append("g")
      .attr("class", "legend")
      .attr("transform", "translate(10,-10)");

  // Create a legend using the data
  // Each text element's offset is computed on the fly using
  // the previous siblings' widths.
  // Text colors are obtained using the color ordinal scale.
  var sumOffsets = 0;
      .text(function(d) { return d.key; })
      .attr("transform", function(d) {
        var offset = this.previousSibling ? this.previousSibling.getBBox().width : 0;
        sumOffsets += offset + 10;
        return "translate(" + sumOffsets + ",0)";
      .attr("fill", function(d) { return color(d.key); });