Related
I will attempt to explain my issue as clearly as possible while also avoid making this topic too long. I recently found the Chart.js library, which is excellent for what I need. Now, since I am using Node.js and need a png of the graph, I am utilizing the chartjs-node-canvas library. Having this information in mind, I will try to split my topic into multiple sections for a clearer understanding.
Ultimate Goal
Before getting into the problem itself, I would like to discuss my ultimate goal. This is to give a general idea on what I'm trying to do so the responses are fitted accordingly. To keep this short, I have data in the form of {awardedDate: "2022-06-22T12:21:17.22Z", badgeId: 1234567}, with awardedDate being a timestamp of when the badge was awarded, and the badgeId being the ID of the badge that was awarded (which is irrelevant to the graph, but it exists because it's part of the data). Now, I have a sample with around 2,787 of these objects, with all having different award dates and IDs, and with dates ranging from 2016 to 2022. My objective is to group these badges by month-year, and that month-year will have the amount of badges earned for that month during that year. With that data, I then want to make a waterfall graph which is based on the amount of badges earned that month of that year. As of right now, there isn't a specific structure on how this will look like, but it could range from an object that looks like {"02-2022": 10, "03-2022": 5} to anything else. I can of course restructure this format based on what is required for a waterfall graph.
Actual Questions
Now that you have a general idea of what my ultimate goal is, my actual question is how I'd be able to make a floating (we can leave the waterfall structure stuff for another topic) bar graph with that data. Since the data can have blank periods (it is possible for a dataset to have gaps that are months long), I cannot really utilize labels (unless I am saying something wrong), so an x-y relation works the best. I tried using the structure of {x: "2022-06-22T12:21:17.226Z", y: [10, 15]}, but that didn't really yield any results. As of right now, I am using a sample code to test how the graph reacts with the data, and of course I'll replace the test values with actual values once I have a finished product. Here is my code so far:
const config = {
type: "bar",
data: {
datasets: [{
label: "Badges",
data: [
{
x: "2022-06-22T12:41:17.226Z",
y: [10, 15]
}
],
borderColor: "rgb(75, 192, 192)",
borderSkipped: false
}]
},
options: {
plugins: {
legend: {
display: false
},
title: {
display: true,
text: "Test",
color: "#FFFFFF"
}
},
scales: {
x: {
type: 'time',
title: {
display: true,
text: 'Time',
color: "#FFFFFF"
},
min: "2022-06-22T12:21:17.226Z",
max: "2022-06-22T14:21:17.226Z",
grid: {
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
ticks: {
color: "#FFFFFF"
}
},
y: {
title: {
display: true,
text: 'Number of Badges',
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
min: 0,
max: 50,
grid: {
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
ticks: {
color: "#FFFFFF"
}
}
}
},
plugins: [
{
id: 'custom_canvas_background_color',
beforeDraw: (chart) => {
const ctx = chart.ctx;
ctx.save();
ctx.fillStyle = '#303030';
ctx.fillRect(0, 0, chart.width, chart.height);
ctx.restore();
}
}
]
};
const imageBuffer = await canvasRenderService.renderToBuffer(config)
fs.writeFileSync("./chart2.png", imageBuffer)
And this is the graph that the code produces:
What is supposed to happen, of course, is that a float bar should be generated near the start that ranges from 5 to 10, but as seen above, nothing happens. If someone could assist me in my problem, that would be amazing. Thank you very much for your time and help, I greatly appreciate it.
Inspired by this answer, I came up with the following solution.
const baseData = [
{ awardedDate: "2022-06-22T12:21:17.22Z" },
{ awardedDate: "2022-06-18T12:21:17.22Z" },
{ awardedDate: "2022-06-15T12:21:17.22Z" },
{ awardedDate: "2022-05-20T12:21:17.22Z" },
{ awardedDate: "2022-05-10T12:21:17.22Z" },
{ awardedDate: "2022-04-16T12:21:17.22Z" },
{ awardedDate: "2022-04-09T12:21:17.22Z" },
{ awardedDate: "2022-04-03T12:21:17.22Z" },
{ awardedDate: "2022-04-01T12:21:17.22Z" },
{ awardedDate: "2022-02-18T12:21:17.22Z" },
{ awardedDate: "2022-02-12T12:21:17.22Z" },
{ awardedDate: "2022-01-17T12:21:17.22Z" }
];
const badgesPerMonth = baseData
.map(o => o.awardedDate)
.sort()
.map(v => moment(v))
.map(m => m.format('MMM YYYY'))
.reduce((acc, month) => {
const badges = acc[month] || 0;
acc[month] = badges + 1;
return acc;
}, {});
const months = Object.keys(badgesPerMonth);
const labels = months.concat('Total');
const data = [];
let total = 0;
for (let i = 0; i < months.length; i++) {
const vStart = total;
total += badgesPerMonth[months[i]];
data.push([vStart, total]);
}
data.push(total);
const backgroundColors = data
.map((o, i) => 'rgba(255, 99, 132, ' + (i + (11 - data.length)) * 0.1 + ')');
new Chart('badges', {
type: 'bar',
data: {
labels: labels,
datasets: [{
label: 'Badges',
data: data,
backgroundColor: backgroundColors,
barPercentage: 1,
categoryPercentage: 0.95
}]
},
options: {
plugins: {
tooltip: {
callbacks: {
label: ctx => {
const v = data[ctx.dataIndex];
return Array.isArray(v) ? v[1] - v[0] : v;
}
}
}
},
scales: {
y: {
ticks: {
beginAtZero: true,
stepSize: 2
}
}
}
}
});
<script src="https://rawgit.com/moment/moment/2.2.1/min/moment.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/Chart.js/3.8.0/chart.min.js"></script>
<canvas id="badges" height="95"></canvas>
If you also want to see the gaps, you would first have to initialize badgesPerMonth with following months between the earliest and latest date, each with value zero. Please take a look at this answer to get an idea about how this could be done.
After reading #uminder's reply, I was able to create the following code which solved my problem:
dateGroups = Object.fromEntries(
Object.entries(dateGroups).sort(([d1,],[d2,]) => {return (d1 < d2) ? -1 : ((d1 > d2) ? 1 : 0)})
)
const dateTimesConst = Object.keys(dateGroups)
const dateValuesConst = Object.values(dateGroups)
let dateTimes = []
let dateValues = []
let prevLength = 0
let mostBadgesPerMonth = 0
for (let i = 0; i < dateValuesConst.length; i++) {
const currentMonth = new Date(Date.parse(dateTimesConst[i]))
const previousMonth = new Date(Date.UTC(currentMonth.getUTCFullYear(), currentMonth.getUTCMonth() - 1, 1, 0, 0, 0, 0)).toISOString()
const nextMonth = new Date(Date.UTC(currentMonth.getUTCFullYear(), currentMonth.getUTCMonth() + 1, 1, 0, 0, 0, 0)).toISOString()
// if (!dateTimesConst.includes(previousMonth)) prevLength = 0
const length = dateValuesConst[i].length
dateValues.push([prevLength, length])
dateTimes.push(dateTimesConst[i])
prevLength = length
if (length > mostBadgesPerMonth) mostBadgesPerMonth = length
// if (!dateTimesConst.includes(nextMonth) && i !== dateValuesConst.length - 1) {
// dateTimes.push(nextMonth)
// dateValues.push([length, 0])
// prevLength = 0
// }
}
function barColorCode() {
return (ctx) => {
const start = ctx.parsed._custom.start
const end = ctx.parsed._custom.end
return start <= end ? "rgba(50, 168, 82, 1)" : (start > end) ? "rgba(191, 27, 27, 1)" : "black"
}
}
const config = {
type: "bar",
data: {
labels: dateTimes,
datasets: [{
label: "Badges",
data: dateValues,
elements: {
bar: {
backgroundColor: barColorCode()
}
},
barPercentage: 1,
categoryPercentage: 0.95,
borderSkipped: false
}]
},
options: {
plugins: {
legend: {
display: false
},
title: {
display: true,
text: "Test",
color: "#FFFFFF"
}
},
scales: {
x: {
type: 'time',
title: {
display: true,
text: 'Date',
color: "#FFFFFF"
},
time: {
unit: "month",
round: "month"
},
min: dateTimesConst[0],
max: dateTimesConst[dateTimesConst.length - 1],
grid: {
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
ticks: {
color: "#FFFFFF"
}
},
y: {
title: {
display: true,
text: 'Number of Badges',
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
min: 0,
max: mostBadgesPerMonth + 1,
grid: {
borderColor: "#FFFFFF",
color: "#FFFFFF"
},
ticks: {
color: "#FFFFFF"
}
}
}
},
plugins: [
{
id: 'custom_canvas_background_color',
beforeDraw: (chart) => {
const ctx = chart.ctx;
ctx.save();
ctx.fillStyle = '#303030';
ctx.fillRect(0, 0, chart.width, chart.height);
ctx.restore();
}
}
]
};
const imageBuffer = await canvasRenderService.renderToBuffer(config)
fs.writeFileSync("./chart2.png", imageBuffer)
Again, big thanks to #uminder for the inspiration.
I am stuck with a problem on chart js while creating line chart. I want to create a chart with the specified data and also need to have horizontal and vertical line while I hover on intersection point. I am able to create vertical line on hover but can not find any solution where I can draw both the line. Here is my code to draw vertical line on hover.
window.lineOnHover = function(){
Chart.defaults.LineWithLine = Chart.defaults.line;
Chart.controllers.LineWithLine = Chart.controllers.line.extend({
draw: function(ease) {
Chart.controllers.line.prototype.draw.call(this, ease);
if (this.chart.tooltip._active && this.chart.tooltip._active.length) {
var activePoint = this.chart.tooltip._active[0],
ctx = this.chart.ctx,
x = activePoint.tooltipPosition().x,
topY = this.chart.legend.bottom,
bottomY = this.chart.chartArea.bottom;
// draw line
ctx.save();
ctx.beginPath();
ctx.moveTo(x, topY);
ctx.lineTo(x, bottomY);
ctx.lineWidth = 1;
ctx.setLineDash([3,3]);
ctx.strokeStyle = '#FF4949';
ctx.stroke();
ctx.restore();
}
}
});
}
//create chart
var backhaul_wan_mos_chart = new Chart(backhaul_wan_mos_chart, {
type: 'LineWithLine',
data: {
labels: ['Aug 1', 'Aug 2', 'Aug 3', 'Aug 4', 'Aug 5', 'Aug 6', 'Aug 7', 'Aug 8'],
datasets: [{
label: 'Series 1',
data: [15, 16, 17, 18, 16, 18, 17, 14, 19, 16, 15, 15, 17],
pointRadius: 0,
fill: false,
borderDash: [3, 3],
borderColor: '#0F1731',
// backgroundColor: '#FF9CE9',
// pointBackgroundColor: ['#FB7BDF'],
borderWidth: 1
}],
// lineAtIndex: 2,
},
options: {
tooltips: {
intersect: false
},
legend: {
display: false
},
scales: {
xAxes: [{
gridLines: {
offsetGridLines: true
},
ticks: {
fontColor: '#878B98',
fontStyle: "600",
fontSize: 10,
fontFamily: "Poppins"
}
}],
yAxes: [{
display: true,
stacked: true,
ticks: {
min: 0,
max: 50,
stepSize: 10,
fontColor: '#878B98',
fontStyle: "500",
fontSize: 10,
fontFamily: "Poppins"
}
}]
},
responsive: true,
}
});
my output of the code is as follow in WAN MoS Score graph --
So I want to have an horizontal line with the same vertical line together when I hover on the intersection (plotted) point..
Please help my guys..Thanks in advance.
You can just add a second draw block for the y coordinate that you get from the tooltip, first you move to the left of the chartArea that you can get the same way you got bottom and top and then you move to the right on the same Y
Chart.defaults.LineWithLine = Chart.defaults.line;
Chart.controllers.LineWithLine = Chart.controllers.line.extend({
draw: function(ease) {
Chart.controllers.line.prototype.draw.call(this, ease);
if (this.chart.tooltip._active && this.chart.tooltip._active.length) {
var activePoint = this.chart.tooltip._active[0],
ctx = this.chart.ctx,
x = activePoint.tooltipPosition().x,
y = activePoint.tooltipPosition().y,
topY = this.chart.legend.bottom,
bottomY = this.chart.chartArea.bottom,
left = this.chart.chartArea.left,
right = this.chart.chartArea.right;
// Set line opts
ctx.save();
ctx.lineWidth = 1;
ctx.setLineDash([3, 3]);
ctx.strokeStyle = '#FF4949';
// draw vertical line
ctx.beginPath();
ctx.moveTo(x, topY);
ctx.lineTo(x, bottomY);
ctx.stroke();
// Draw horizontal line
ctx.beginPath();
ctx.moveTo(left, y);
ctx.lineTo(right, y);
ctx.stroke();
ctx.restore();
}
}
});
var options = {
type: 'LineWithLine',
data: {
labels: ["Red", "Blue", "Yellow", "Green", "Purple", "Orange"],
datasets: [{
label: '# of Votes',
data: [12, 19, 3, 5, 2, 3],
borderWidth: 1
},
{
label: '# of Points',
data: [7, 11, 5, 8, 3, 7],
borderWidth: 1
}
]
},
options: {
}
}
var ctx = document.getElementById('chartJSContainer').getContext('2d');
new Chart(ctx, options);
<body>
<canvas id="chartJSContainer" width="600" height="400"></canvas>
<script src="https://cdnjs.cloudflare.com/ajax/libs/Chart.js/2.9.4/Chart.js"></script>
</body>
Edit:
You should use a custom plugin for this since you dont draw everytime you move the cursor and you can enforce this by using a custom plugin:
const options = {
type: 'line',
data: {
labels: ["Red", "Blue", "Yellow", "Green", "Purple", "Orange"],
datasets: [{
label: '# of Votes',
data: [12, 19, 3, 5, 2, 3],
borderWidth: 1
},
{
label: '# of Points',
data: [7, 11, 5, 8, 3, 7],
borderWidth: 1
}
]
},
options: {
plugins: {
corsair: {
dash: [2, 2],
color: 'red',
width: 3
}
}
},
plugins: [{
id: 'corsair',
afterInit: (chart) => {
chart.corsair = {
x: 0,
y: 0
}
},
afterEvent: (chart, evt) => {
const {
chartArea: {
top,
bottom,
left,
right
}
} = chart;
const {
x,
y
} = evt;
if (x < left || x > right || y < top || y > bottom) {
chart.corsair = {
x,
y,
draw: false
}
chart.draw();
return;
}
chart.corsair = {
x,
y,
draw: true
}
chart.draw();
},
afterDatasetsDraw: (chart, _, opts) => {
const {
ctx,
chartArea: {
top,
bottom,
left,
right
}
} = chart;
const {
x,
y,
draw
} = chart.corsair;
if (!draw) {
return;
}
ctx.lineWidth = opts.width || 0;
ctx.setLineDash(opts.dash || []);
ctx.strokeStyle = opts.color || 'black'
ctx.save();
ctx.beginPath();
ctx.moveTo(x, bottom);
ctx.lineTo(x, top);
ctx.moveTo(left, y);
ctx.lineTo(right, y);
ctx.stroke();
ctx.restore();
}
}]
}
const ctx = document.getElementById('chartJSContainer').getContext('2d');
new Chart(ctx, options);
<body>
<canvas id="chartJSContainer" width="600" height="400"></canvas>
<script src="https://cdnjs.cloudflare.com/ajax/libs/Chart.js/2.9.4/Chart.js"></script>
</body>
Edit:
Updated answer for v3
const options = {
type: 'line',
data: {
labels: ["Red", "Blue", "Yellow", "Green", "Purple", "Orange"],
datasets: [{
label: '# of Votes',
data: [12, 19, 3, 5, 2, 3],
borderWidth: 1
},
{
label: '# of Points',
data: [7, 11, 5, 8, 3, 7],
borderWidth: 1
}
]
},
options: {
plugins: {
corsair: {
dash: [2, 2],
color: 'red',
width: 3
}
}
},
plugins: [{
id: 'corsair',
afterInit: (chart) => {
chart.corsair = {
x: 0,
y: 0
}
},
afterEvent: (chart, evt) => {
const {
chartArea: {
top,
bottom,
left,
right
}
} = chart;
const {
event: {
x,
y
}
} = evt;
if (x < left || x > right || y < top || y > bottom) {
chart.corsair = {
x,
y,
draw: false
}
chart.draw();
return;
}
chart.corsair = {
x,
y,
draw: true
}
chart.draw();
},
afterDatasetsDraw: (chart, _, opts) => {
const {
ctx,
chartArea: {
top,
bottom,
left,
right
}
} = chart;
const {
x,
y,
draw
} = chart.corsair;
if (!draw) {
return;
}
ctx.lineWidth = opts.width || 0;
ctx.setLineDash(opts.dash || []);
ctx.strokeStyle = opts.color || 'black'
ctx.save();
ctx.beginPath();
ctx.moveTo(x, bottom);
ctx.lineTo(x, top);
ctx.moveTo(left, y);
ctx.lineTo(right, y);
ctx.stroke();
ctx.restore();
}
}]
}
const ctx = document.getElementById('chartJSContainer').getContext('2d');
new Chart(ctx, options);
<body>
<canvas id="chartJSContainer" width="600" height="400"></canvas>
<script src="https://cdnjs.cloudflare.com/ajax/libs/Chart.js/3.8.0/chart.js"></script>
</body>
This is 2022, the current version of ChartJS is 4.0.1. So, I recommend to use this new implementation.
First, let's define a plugin. ChartJS's plugins has an id parameter, in this case less say corsair.
Then we define default variables for our plugin, like width, color and line dash. Additionally, our plugin will have three parameters: x, y, and draw. x and y are the values of the mousemove event and draw represents the inChartArea parameter, this parameter defines if the event occurred inside of the chart area or not.
Finally, we capture the afterDraw hook to draw a vertical and horizontal lines based on the x and y values if the event was triggered inside of the chart area.
ChartJS has various hooks to capture different parts of the chart render cycle.
const plugin = {
id: 'corsair',
defaults: {
width: 1,
color: '#FF4949',
dash: [3, 3],
},
afterInit: (chart, args, opts) => {
chart.corsair = {
x: 0,
y: 0,
}
},
afterEvent: (chart, args) => {
const {inChartArea} = args
const {type,x,y} = args.event
chart.corsair = {x, y, draw: inChartArea}
chart.draw()
},
beforeDatasetsDraw: (chart, args, opts) => {
const {ctx} = chart
const {top, bottom, left, right} = chart.chartArea
const {x, y, draw} = chart.corsair
if (!draw) return
ctx.save()
ctx.beginPath()
ctx.lineWidth = opts.width
ctx.strokeStyle = opts.color
ctx.setLineDash(opts.dash)
ctx.moveTo(x, bottom)
ctx.lineTo(x, top)
ctx.moveTo(left, y)
ctx.lineTo(right, y)
ctx.stroke()
ctx.restore()
}
}
const data = {
labels: ["Red", "Blue", "Yellow", "Green", "Purple", "Orange"],
datasets: [{
label: '# of Votes',
data: [12, 19, 3, 5, 2, 3],
borderWidth: 1
},
{
label: '# of Points',
data: [7, 11, 5, 8, 3, 7],
borderWidth: 1
}
]
}
const options = {
maintainAspectRatio: false,
hover: {
mode: 'index',
intersect: false,
},
plugins: {
corsair: {
color: 'black',
}
}
}
const config = {
type: 'line',
data,
options,
plugins: [plugin],
}
const $chart = document.getElementById('chart')
const chart = new Chart($chart, config)
<div class="wrapper" style="width: 98vw; height: 180px">
<canvas id="chart"></canvas>
</div>
<script src="https://unpkg.com/chart.js#4.0.1/dist/chart.umd.js"></script>
I have done exactly this (but vertical line only) in a previous version of one of my projects. Unfortunately this feature has been removed but the older source code file can still be accessed via my github.
The key is this section of the code:
Chart.defaults.LineWithLine = Chart.defaults.line;
Chart.controllers.LineWithLine = Chart.controllers.line.extend({
draw: function(ease) {
Chart.controllers.line.prototype.draw.call(this, ease);
if (this.chart.tooltip._active && this.chart.tooltip._active.length) {
var activePoint = this.chart.tooltip._active[0],
ctx = this.chart.ctx,
x = activePoint.tooltipPosition().x,
topY = this.chart.legend.bottom,
bottomY = this.chart.chartArea.bottom;
// draw line
ctx.save();
ctx.beginPath();
ctx.moveTo(x, topY);
ctx.lineTo(x, bottomY);
ctx.lineWidth = 0.5;
ctx.strokeStyle = '#A6A6A6';
ctx.stroke();
ctx.restore();
}
}
});
Another caveat is that the above code works with Chart.js 2.8 and I am aware that the current version of Chart.js is 3.1. I haven't read the official manual on the update but my personal experience is that this update is not 100% backward-compatible--so not sure if it still works if you need Chart.js 3. (But sure you may try 2.8 first and if it works you can then somehow tweak the code to make it work on 3.1)
I'm creating some pie charts with Plotly.js but I can't figure out how to set the value format. If my number is 19.231, I get 19.2. I want 19.2310, with 4 digit precision. My data and layout follow here:
var data = [{
values: values,
labels: labels,
type: 'pie',
marker: {
colors: colors
},
textfont: {
family: 'Helvetica, sans-serif',
size: 48,
color: '#000'
}
}];
var layout = {
height: 1350,
width: 1500,
title: title,
titlefont: {
family: 'Helvetica, sans-serif',
size: 58,
color: '#000'
},
legend: {
x: 1.1,
y: 0.5,
size: 40,
font: {
family: 'Helvetica, sans-serif',
size: 48,
color: '#000'
}
}
};
As far as I know you cannot do that directly in Plotly but you could add text values with your desired precision and set textinfo to text.
var values = [Math.random(), Math.random(), Math.random()];
var sum = values.reduce(function(pv, cv) { return pv + cv; }, 0);
var digits = 4;
var rounded_values = [];
for (var i = 0; i < values.length; i += 1) {
rounded_values.push(Math.round(values[i]/sum * 100 * 10**digits) / 10**digits + '%');
}
var myPlot = document.getElementById('myPlot');
var data = [{
values: values,
type: 'pie',
text: rounded_values,
textinfo: 'text'
}];
Plotly.plot(myPlot, data);
<script src="https://cdn.plot.ly/plotly-latest.min.js"></script>
<div id="myPlot"></div>
I saw the document find some rule
about Advanced Hovertemplate it said
data: [
{
type: "scatter",
mode: "markers",
x: x,
y: y,
text: t,
marker: { size: s, sizeref: 4000, sizemode: "area" },
transforms: [{ type: "groupby", groups: c }],
hovertemplate:
"<b>%{text}</b><br><br>" +
"%{yaxis.title.text}: %{y:$,.0f}<br>" +
"%{xaxis.title.text}: %{x:.0%}<br>" +
"Number Employed: %{marker.size:,}" +
"<extra></extra>"
}
],
you can change here %{y:$,.0f} 0f to 4f .
like this
hovertemplate: '%{y:$,.4f}',
Polt.ly JS document
I've been trying to display somewhat complex data on my webpage and chose chart.js to do so.
Therefor I need to group multiple stacked bars horizontally.
I already found this fiddle for "normal" bars but couldn't quite change it to work with horizontalBar yet.
Stackoverflow question: Chart.js stacked and grouped bar chart
The original Fiddle (http://jsfiddle.net/2xjwoLq0/) has
Chart.defaults.groupableBar = Chart.helpers.clone(Chart.defaults.bar);
And I just replaced the .bar everywhere in the code with .horizontalBar (well knowing that this won't make the cut).
Chart.defaults.groupableBar = Chart.helpers.clone(Chart.defaults.horizontalBar);
Since that didn't quite work, I tried adding the second stacked modifier as suggested for horizontal bars here:
Horizontal stacked bar chart with chart.js and flipped the functions for X and Y calculation (calculateBarY/calculateBarX)
Which quite work either because the stacks won't get merged onto each other correctly.
http://jsfiddle.net/2xjwoLq0/3/
I would appreciate if anyone could help me out on this one.
Looking for something similar, I took a look on example you gave, and decide to write something.
Rather than trying to fix the code or reusing the 'groupableBar', I get Chart.js code from Chart.controllers.horizontalBar and rewrite some part in functions calculateBarY, calculateBarHeight.
Just reused the getBarCount function from your example.
Chart.defaults.groupableHBar = Chart.helpers.clone(Chart.defaults.horizontalBar);
Chart.controllers.groupableHBar = Chart.controllers.horizontalBar.extend({
calculateBarY: function(index, datasetIndex, ruler) {
var me = this;
var meta = me.getMeta();
var yScale = me.getScaleForId(meta.yAxisID);
var barIndex = me.getBarIndex(datasetIndex);
var topTick = yScale.getPixelForValue(null, index, datasetIndex, me.chart.isCombo);
topTick -= me.chart.isCombo ? (ruler.tickHeight / 2) : 0;
var stackIndex = this.getMeta().stackIndex;
if (yScale.options.stacked) {
if(ruler.datasetCount>1) {
var spBar=ruler.categorySpacing/ruler.datasetCount;
var h=me.calculateBarHeight(ruler);
return topTick + (((ruler.categoryHeight - h) / 2)+ruler.categorySpacing-spBar/2)+(h+spBar)*stackIndex;
}
return topTick + (ruler.categoryHeight / 2) + ruler.categorySpacing;
}
return topTick +
(ruler.barHeight / 2) +
ruler.categorySpacing +
(ruler.barHeight * barIndex) +
(ruler.barSpacing / 2) +
(ruler.barSpacing * barIndex);
},
calculateBarHeight: function(ruler) {
var returned=0;
var me = this;
var yScale = me.getScaleForId(me.getMeta().yAxisID);
if (yScale.options.barThickness) {
returned = yScale.options.barThickness;
}
else {
returned= yScale.options.stacked ? ruler.categoryHeight : ruler.barHeight;
}
if(ruler.datasetCount>1) {
returned=returned/ruler.datasetCount;
}
return returned;
},
getBarCount: function () {
var stacks = [];
// put the stack index in the dataset meta
Chart.helpers.each(this.chart.data.datasets, function (dataset, datasetIndex) {
var meta = this.chart.getDatasetMeta(datasetIndex);
if (meta.bar && this.chart.isDatasetVisible(datasetIndex)) {
var stackIndex = stacks.indexOf(dataset.stack);
if (stackIndex === -1) {
stackIndex = stacks.length;
stacks.push(dataset.stack);
}
meta.stackIndex = stackIndex;
}
}, this);
this.getMeta().stacks = stacks;
return stacks.length;
}
});
var data = {
labels: ["January", "February", "March"],
datasets: [
{
label: "Dogs",
backgroundColor: "rgba(255,0,0,0.2)",
data: [20, 10, 25],
stack: 1,
xAxisID: 'x-axis-0',
yAxisID: 'y-axis-0'
},
{
label: "Cats",
backgroundColor: "rgba(255,255,0,0.2)",
data: [70, 85, 65],
stack: 1,
xAxisID: 'x-axis-0',
yAxisID: 'y-axis-0'
},
{
label: "Birds",
backgroundColor: "rgba(0,255,255,0.2)",
data: [10, 5, 10],
stack: 1,
xAxisID: 'x-axis-0',
yAxisID: 'y-axis-0'
},
{
label: ":-)",
backgroundColor: "rgba(0,255,0,0.2)",
data: [20, 10, 30],
stack: 2,
xAxisID: 'x-axis-1',
yAxisID: 'y-axis-0'
},
{
label: ":-|",
backgroundColor: "rgba(0,0,255,0.2)",
data: [40, 50, 20],
stack: 2,
xAxisID: 'x-axis-1',
yAxisID: 'y-axis-0'
},
{
label: ":-(",
backgroundColor: "rgba(0,0,0,0.2)",
data: [60, 20, 20],
stack: 2,
xAxisID: 'x-axis-1',
yAxisID: 'y-axis-0'
},
]
};
var ctx = document.getElementById("myChart").getContext("2d");
new Chart(ctx, {
type: 'groupableHBar',
data: data,
options: {
scales: {
yAxes: [{
stacked: true,
type: 'category',
id: 'y-axis-0'
}],
xAxes: [{
stacked: true,
type: 'linear',
ticks: {
beginAtZero:true
},
gridLines: {
display: false,
drawTicks: true,
},
id: 'x-axis-0'
},
{
stacked: true,
position: 'top',
type: 'linear',
ticks: {
beginAtZero:true
},
id: 'x-axis-1',
gridLines: {
display: true,
drawTicks: true,
},
display: false
}]
}
}
});
<script src="https://cdnjs.cloudflare.com/ajax/libs/Chart.js/2.4.0/Chart.min.js"></script>
<canvas id="myChart"></canvas>
Also put example on jsfiddle here: https://jsfiddle.net/b7gnron7/4/
Code is not strongly tested, you might found some bugs especially if you try to display only one stacked group (use horizontalBar instead in this case).
Your post is a little bit old... not sure that you still need a solution, but it could be useful for others ^_^
With a bar chart like this one, is is possible to change the width of the bars to represent another data attribute, say the weight of the fruits. The heavier the fruit is, the thicker the bar.
You play with the script here. I am open to other javascript plotting libraries that could do that as long as they are free.
$(function () {
var chart;
$(document).ready(function() {
chart = new Highcharts.Chart({
chart: {
renderTo: 'container',
type: 'column'
},
title: {
text: 'Column chart with negative values'
},
xAxis: {
categories: ['Apples', 'Oranges', 'Pears', 'Grapes', 'Bananas']
},
tooltip: {
formatter: function() {
return ''+
this.series.name +': '+ this.y +'';
}
},
credits: {
enabled: false
},
series: [{
name: 'John',
data: [5, 3, 4, 7, 2]
// I would like something like this (3.5, 6 etc is the width) :
// data: [[5, 3.4], [3, 6], [4, 3.4], [7, 2], [2, 5]]
}, {
name: 'Jane',
data: [2, -2, -3, 2, 1]
}, {
name: 'Joe',
data: [3, 4, 4, -2, 5]
}]
});
});
});
pointWidth is what you require to set the width of the bars. try
plotOptions: {
series: {
pointWidth: 15
}
}
This display bars with the width of 15px. Play around here. Just made an edit to the already existing code.
I use a set of area charts to simulate a variable-width-column/bar-chart. Say, each column/bar is represented by a rectangle area.
See my fiddle demo (http://jsfiddle.net/calfzhou/TUt2U/).
$(function () {
var rawData = [
{ name: 'A', x: 5.2, y: 5.6 },
{ name: 'B', x: 3.9, y: 10.1 },
{ name: 'C', x: 11.5, y: 1.2 },
{ name: 'D', x: 2.4, y: 17.8 },
{ name: 'E', x: 8.1, y: 8.4 }
];
function makeSeries(listOfData) {
var sumX = 0.0;
for (var i = 0; i < listOfData.length; i++) {
sumX += listOfData[i].x;
}
var gap = sumX / rawData.length * 0.2;
var allSeries = []
var x = 0.0;
for (var i = 0; i < listOfData.length; i++) {
var data = listOfData[i];
allSeries[i] = {
name: data.name,
data: [
[x, 0], [x, data.y],
{
x: x + data.x / 2.0,
y: data.y,
dataLabels: { enabled: true, format: data.x + ' x {y}' }
},
[x + data.x, data.y], [x + data.x, 0]
],
w: data.x,
h: data.y
};
x += data.x + gap;
}
return allSeries;
}
$('#container').highcharts({
chart: { type: 'area' },
xAxis: {
tickLength: 0,
labels: { enabled: false}
},
yAxis: {
title: { enabled: false}
},
plotOptions: {
area: {
marker: {
enabled: false,
states: {
hover: { enabled: false }
}
}
}
},
tooltip: {
followPointer: true,
useHTML: true,
headerFormat: '<span style="color: {series.color}">{series.name}</span>: ',
pointFormat: '<span>{series.options.w} x {series.options.h}</span>'
},
series: makeSeries(rawData)
});
});
Fusioncharts probably is the best option if you have a license for it to do the more optimal Marimekko charts…
I've done a little work trying to get a Marimekko charts solution in highcharts. It's not perfect, but approximates the first Marimekko charts example found here on the Fusion Charts page…
http://www.fusioncharts.com/resources/chart-tutorials/understanding-the-marimekko-chart/
The key is to use a dateTime axis, as that mode provides you more flexibility for the how you distribute points and line on the X axis which provides you the ability to have variably sized "bars" that you can construct on this axis. I use 0-1000 second space and outside the chart figure out the mappings to this scale to approximate percentage values to pace your vertical lines. Here ( http://jsfiddle.net/miken/598d9/2/ ) is a jsfiddle example that creates a variable width column chart.
$(function () {
var chart;
Highcharts.setOptions({
colors: [ '#75FFFF', '#55CCDD', '#60DD60' ]
});
$(document).ready(function() {
var CATEGORY = { // number out of 1000
0: '',
475: 'Desktops',
763: 'Laptops',
1000: 'Tablets'
};
var BucketSize = {
0: 475,
475: 475,
763: 288,
1000: 237
};
chart = new Highcharts.Chart({
chart: {
renderTo: 'container',
type: 'area'
},
title: {
text: 'Contribution to Overall Sales by Brand & Category (in US$)<br>(2011-12)'
},
xAxis: {
min: 0,
max: 1000,
title: {
text: '<b>CATEGORY</b>'
},
tickInterval: 1,
minTickInterval: 1,
dateTimeLabelFormats: {
month: '%b'
},
labels: {
rotation: -60,
align: 'right',
formatter: function() {
if (CATEGORY[this.value] !== undefined) {
return '<b>' + CATEGORY[this.value] + ' (' +
this.value/10 + '%)</b>';
}
}
}
},
yAxis: {
max: 100,
gridLineWidth: 0,
title: {
text: '<b>% Share</b>'
},
labels: {
formatter: function() {
return this.value +'%'
}
}
},
tooltip: {
shared: true,
useHTML: true,
formatter: function () {
var result = 'CATEGORY: <b>' +
CATEGORY[this.x] + ' (' + Highcharts.numberFormat(BucketSize[this.x]/10,1) + '% sized bucket)</b><br>';
$.each(this.points, function(i, datum) {
if (datum.point.y !== 0) {
result += '<span style="color:' +
datum.series.color + '"><b>' +
datum.series.name + '</b></span>: ' +
'<b>$' + datum.point.y + 'K</b> (' +
Highcharts.numberFormat(
datum.point.percentage,2) +
'%)<br/>';
}
});
return (result);
}
},
plotOptions: {
area: {
stacking: 'percent',
lineColor: 'black',
lineWidth: 1,
marker: {
enabled: false
},
step: true
}
},
legend: {
layout: 'vertical',
align: 'right',
verticalAlign: 'top',
x: 0,
y: 100,
borderWidth: 1,
title: {
text : 'Brand:'
}
},
series: [ {
name: 'HP',
data: [
[0,298],
[475,109],
[763,153],
[1000,153]
]
}, {
name: 'Dell',
data: [
[0,245],
[475,198],
[763,120],
[1000,120]
]
}, {
name: 'Sony',
data: [
[0,335],
[475,225],
[763,164],
[1000,164]
]
}]
},
function(chart){
// Render bottom line.
chart.renderer.path(['M', chart.plotLeft, chart.plotHeight + 66, 'L', chart.plotLeft+chart.plotWidth, chart.plotHeight + 66])
.attr({
'stroke-width': 3,
stroke: 'black',
zIndex:50
})
.add();
for (var category_idx in CATEGORY) {
chart.renderer.path(['M', (Math.round((category_idx / 1000) * chart.plotWidth)) + chart.plotLeft, 66, 'V', chart.plotTop + chart.plotHeight])
.attr({
'stroke-width': 1,
stroke: 'black',
zIndex:4
})
.add();
}
});
});
});
It adds an additional array to allow you to map category names to second tic values to give you a more "category" view that you might want. I've also added code at the bottom that adds vertical dividing lines between the different columns and the bottom line of the chart. It might need some tweaks for the size of your surrounding labels, etc. that I've hardcoded in pixels here as part of the math, but it should be doable.
Using a 'percent' type accent lets you have the y scale figure out the percentage totals from the raw data, whereas as noted you need to do your own math for the x axis. I'm relying more on a tooltip function to provide labels, etc than labels on the chart itself.
Another big improvement on this effort would be to find a way to make the tooltip hover area and labels to focus and be centered and encompass the bar itself instead of the right border of each bar that it is now. If someone wants to add that, feel free to here.
If I got it right you want every single bar to be of different width. I had same problem and struggled a lot to find a library offering this option. I came to the conclusion - there's none.
Anyways, I played with highcharts a little, got creative and came up with this:
You mentioned that you'd like your data to look something like this: data: [[5, 3.4], [3, 6], [4, 3.4]], with the first value being the height and the second being the width.
Let's do it using the highcharts' column graph.
Step 1:
To better differentiate the bars, input each bar as a new series. Since I generated my data dynamically, I had to assign new series dynamically:
const objects: any = [];
const extra = this.data.length - 1;
this.data.map((range) => {
const obj = {
type: 'column',
showInLegend: false,
data: [range[1]],
animation: true,
borderColor: 'black',
borderWidth: 1,
color: 'blue'
};
for (let i = 0; i < extra; i++) {
obj.data.push(null);
}
objects.push(obj);
});
this.chartOptions.series = objects;
That way your different series would look something like this:
series: [{
type: 'column',
data: [5, 3.4]
}, {
type: 'column',
data: [3, 6]
}, {
type: 'column',
data: [4, 3.4]
}]
Step 2:
Assign this as plot options for highcharts:
plotOptions: {
column: {
pointPadding: 0,
borderWidth: 0,
groupPadding: 0,
shadow: false
}
}
Step 3:
Now let's get creative - to have the same starting point for all bars, we need to move every single one to the graph's start:
setColumnsToZero() {
this.data.map((item, index) => {
document.querySelector('.highcharts-series-' + index).children[0].setAttribute('x', '0');
});
}
Step 4:
getDistribution() {
let total = 0;
// Array including all of the bar's data: [[5, 3.4], [3, 6], [4, 3.4]]
this.data.map(item => {
total = total + item[0];
});
// MARK: Get xAxis' total width
const totalWidth = document.querySelector('.highcharts-axis-line').getBoundingClientRect().width;
let pos = 0;
this.data.map((item, index) => {
const start = item[0];
const width = (start * totalWidth) / total;
document.querySelector('.highcharts-series-' + index).children[0].setAttribute('width', width.toString());
document.querySelector('.highcharts-series-' + index).children[0].setAttribute('x', pos.toString());
pos = pos + width;
this.getPointsPosition(index, totalWidth, total);
});
}
Step 4:
Let's get to the xAxis' points. In the first functions modify the already existing points, move the last point to the end of the axis and hide the others. In the second function we clone the last point, modify it to have either 6 or 3 total xAxis points and move each of them to the correct position
getPointsPosition(index, totalWidth, total) {
const col = document.querySelector('.highcharts-series-' + index).children[0];
const point = (document.querySelector('.highcharts-xaxis-labels').children[index] as HTMLElement);
const difference = col.getBoundingClientRect().right - point.getBoundingClientRect().right;
const half = point.getBoundingClientRect().width / 2;
if (index === this.data.length - 1) {
this.cloneNode(point, difference, totalWidth, total);
} else {
point.style.display = 'none';
}
point.style.transform = 'translateX(' + (+difference + +half) + 'px)';
point.innerHTML = total.toString();
}
cloneNode(ref: HTMLElement, difference, totalWidth, total) {
const width = document.documentElement.getBoundingClientRect().width;
const q = total / (width > 1000 && ? 6 : 3);
const w = totalWidth / (width > 1000 ? 6 : 3);
let val = total;
let valW = 0;
for (let i = 0; i < (width > 1000 ? 6 : 3); i++) {
val = val - q;
valW = valW + w;
const clone = (ref.cloneNode(true) as HTMLElement);
document.querySelector('.highcharts-xaxis-labels').appendChild(clone);
const half = clone.getBoundingClientRect().width / 2;
clone.style.transform = 'translateX(' + (-valW + difference + half) + 'px)';
const inner = Math.round(val * 100) / 100;
clone.innerHTML = inner.toString();
}
}
In the end we have a graph looking something like this (not the data from this given example, but for [[20, 0.005], [30, 0.013333333333333334], [20, 0.01], [30, 0.005555555555555555], [20, 0.006666666666666666]] with the first value being the width and the second being the height):
There might be some modifications to do to 100% fit your case. F.e. I had to adjust the xAxis' points a specific starting and end point - I spared this part.