Develop Reference

Chart.js : sparser clickable data points

I've got a chart.js line chart, with 128 datapoints.
Currently, it uses the solution here for drag and drop verticals.
But it's a hassle. I'd like to limit the clickable / editable points to 30 or so.
(Then ideally it would do some smart interpolation of the surrounding values.)
Any ideas how to achieve this?
function move_handler(event)
{
// locate grabbed point in chart data
if (activePoint != null) {
var data = activePoint._chart.data;
var datasetIndex = activePoint._datasetIndex;
// read mouse position
const helpers = Chart.helpers;
var position = helpers.getRelativePosition(event, myChart);
// convert mouse position to chart y axis value
var chartArea = window.myChart.chartArea;
var yAxis = window.myChart.scales["y-axis-0"];
var yValue = map(position.y, chartArea.bottom, chartArea.top, yAxis.min, yAxis.max);
// update y value of active data point
data.datasets[datasetIndex].data[activePoint._index] = yValue;
// try {
// for (let i = -5;i < 5;i++){
// data.datasets[datasetIndex].data[activePoint._index + i] = yValue;
// }
// }
// catch (err) {}
window.myChart.update();
};
};

I found Smooth.js which did the trick

how to pass big data to google scatter chart

I am relatively new to JavaScript and Django and I am struggling with passing big data to my google chart.
I have a chart representing velocities for a given date and distance. In my django views I create list of distances, dates and according velocities. I also generate there a list with sorted values occuring in velocity list and a list with colors according to velocity's value.
I want to have a chart with velocity map with applied colortable like this :
http://i.imgur.com/9Tyv8Rn.jpg
So I used scatter chart with velocity series. The chart is dynamic, it's diffrent for every item selected by a user.
JS to generate rows and columns :
// Define data table rows:
var rows = [];
var rows_list = [];
var vl_max = vel_list.length;
for (i=0; i < vl_max; i+=1) {
var date_tmp = new Date(date_list[i].split(',')[0],date_list[i].split(',')[1]-1,date_list[i].split(',')[2]);
var date = [date_tmp];
var vel_tmp = vel_list[i];
var vtemp_max = vel_tmp.length;
var tooltip_dsname = dsname_list[i];
var tooltip_track = track_list[i];
for (j=0; j < vtemp_max; j+=1) {
var cell = [{v : date_tmp}];
for (k=0; k < vr_max; k+=1) {
var vel_full = vel_tmp[j];
var vel = vel_full.toFixed(1);
if (vel == vel_range[k]) {
// tooltip:
var dist = dist_list[j]/1000;
var yyyy = date_tmp.getFullYear().toString();
var mm = (date_tmp.getMonth()+1).toString(); // getMonth() is zero-based
var dd = date_tmp.getDate().toString();
var tooltip_date = yyyy + "-" + (mm[1]?mm:"0"+mm[0]) + "-" + (dd[1]?dd:"0"+dd[0]);
var tooltip = "<b>dataset: </b>"+tooltip_dsname+"<br><b>date: </b>"+tooltip_date+"<br><b>track: </b>"+tooltip_track+"<br><b>distance: </b>"+dist+" k"+mapunit+"<br><b> velocity: </b>"+vel_full.toFixed(2)+" m/d";
var color = color_list[k]
var style = "point { shape-type: square; fill-color: "+color+";}"
} else {
var dist = NaN;
var tooltip = "empty" ;
var style = "empty" ;
}
cell.push({v: dist},{v: tooltip},{v:style});
}
rows_list.push({c: cell});
}
};
Here is JSfiddle for chart generation with smaller data :
http://jsfiddle.net/joannao89/t26ooyrt/2/
The problem that I have is while the chart is working for smaller data, once I want to load it for a long distance and a wide date range, the browser keeps on popping us this line : "A website is slowing down your browser, what would you like to do ? {stop} {wait}"
I know that this is probably the problem of too large amount of rows, my website generates also 3 other charts like this, with the same data but in another X-Y axis combination (for example time on X-axis, velocity on Y-axis and distance as series) and it works perfectly fine. That's why I would like to pass the data to the chart in some faster way, but I have no clue how.
I already tried to use setTimeout, but it doesn't change a lot. I also tried doing a little less coding on JS side and more in django views, but it also didn't help.
So any suggestions about how to solve this will be very appreciated!

meet strange during build up the basic 2D grids for future simulation in javascript?

I am doing one project while stuck at the first step. The idea is to set up a 3D model. The (x,y) is the geometry area and z is the value of interests. I need simulate all area of (x,y) coordinates.Firstly, I try to initiate the z value for all (x,y) coordinate. I met problems at this step by following codes.
<script type="text/javascript">
// use the rectangle as the "unit" of the 2D model.
var x = [];
var y = [];
var z = [];
var R1 = 5;
var dataset = [];
// use function to initiate a row
var rowGen = function(k){
for (j=0;j<10;j++){
x.push(j*5);
y.push(k*2);
z.push(R1);
dataset.push({x:x[j],y:y[j],z:z[j]});
}
}
// initiate 10 rows
for (i=0;i<10;i++){
rowGen(i);
}
This always just initiate the z value for parts of the (x,y) area. Just the first row (y=0). I checked dataset in console, it has the correct numbers(100) of objects （x,y,z）, while all of them have y=0, which shows in the figure only one line, instead of 2D area.
I tried a different loops to do that but just can not find out the solution. Could you please help bring me some light?

Just found the logic bug. I had pushed the wrong value into y.Changed the code as below works.
// use the rectangle as the "unit" of the 2D model.
var x = [];
var y = [];
var z = [];
var R1 = 5;
var dataset = [];
// use function to initiate a row
var rowGen = function(k){
for (j=0;j<10;j++){
x.push(j*5);
y.push(k*2);
z.push(R1);
dataset.push({x:x[j],y:k*2,z:z[j]});
}
}
// initiate 10 rows
for (i=0;i<10;i++){
rowGen(i);
}

Get nearest all Y values on click of anywhere of linechart using highchart

I have a line chart with multiple trend lines. I want to display the nearest Y values for all trend lines on click of anywhere in chart.
In the screenshot, I want to display all Y values (i.e., displayed in Tooltip) on right side panel corresponding to bitrates on clicking anywhere in chart.
Can I get tooltip data on charts click event?

I have found out a workaround to find all y values for all trend line. But it's expensive as we have to do single iteration over x axis values for a single trend line.
chart:{
events: {
click: function(e){
console.log(getAllYValues(this, e.xAxis[0].value)) //e.xAxis[0].value is the x- coordinate where it got clicked
}
}
}
function getAllYValues(chartObj, xValue) {
var _series0 = chartObj.series[0].points;
var key = getXValue(_series0, xValue); //Get Nearest X value
var yValues = [];
for (var i = 0; i < chartObj.series.length; i++) {
yValues[i].bitsPerSecond = chartObj.series[i].points[key].y
}
return yValues;
}
function getXValue(points, xValue) {
for (var i = 0; i < points.length; i++) {
if (points[i].x >= xValue) {
if(i > 0) {
return Math.abs(points[i].x - xValue) < Math.abs(points[i-1].x - xValue) ? i : i-1; //Get nearest X value
}
else{return i;}
}
}
}

How to set the dynamic or static tick size in a Rickshaw.js plot?

I am creating a Rickshaw.js-powered graph much like in this example: http://code.shutterstock.com/rickshaw/tutorial/example_07.html based on my own data that is returned via an AJAX call. The data is either measured in bytes (typical values range in a few gigabytes or hundreds of MBs) or seconds (anywhere between 10s and 50 minutes). I tried using a Rickshaw.Fixtures.Number.formatBase1024KMGTP formatter for the bytes and wrote my own for the seconds, which does its part well. The problem is that I need to position the tick lines in a smart way - preferably dynamically, but even static settings (e.g. place a tick every 1024*1024*1024=1 GB or every 60 s) would be fine.
I tried setting the tickSize to 1024^3 like so:
var y_axis = new Rickshaw.Graph.Axis.Y({
graph: graph,
tickSize: 1073741824 // 1 GB
});
y_axis.render();
but I ended up seeing no ticks at all. What am I doing wrong and what would be the right way?

Basically, you need to adapt the tickOffsets() function of the Axis.X, Axis.Y and Axis.Time classes in Rickshaw.
tickSize will not help you with that as - like #Old Pro stated correctly - it indicates the size of the bold tick lines in pixels. It has nothing to do with spacing.
For Time-based Axes
My solution essentially consists of replacing the standard tickOffsets() function in those files
this.tickOffsets = function() {
var domain = this.graph.x.domain();
var unit = this.fixedTimeUnit || this.appropriateTimeUnit();
var count = Math.ceil((domain[1] - domain[0]) / unit.seconds);
var runningTick = domain[0];
var offsets = [];
for (var i = 0; i < count; i++) {
var tickValue = time.ceil(runningTick, unit);
runningTick = tickValue + unit.seconds / 2;
offsets.push( { value: tickValue, unit: unit } );
}
return offsets;
};
by a custom routine. This is gonna do the trick:
this.tickOffsets = function() {
var domain = this.graph.x.domain();
var unit = this.fixedTimeUnit || this.appropriateTimeUnit();
var tickSpacing = args.tickSpacing || unit.seconds;
var count = Math.ceil((domain[1] - domain[0]) / tickSpacing);
var runningTick = domain[0];
var offsets = [];
for (var i = 0; i < count; i++) {
var tickValue = time.ceil(runningTick, unit);
runningTick = tickValue + tickSpacing;
offsets.push( { value: tickValue, unit: unit } );
}
return offsets;
};
With that in place, you can write something like
var time = new Rickshaw.Fixtures.Time();
var timeUnit = time.unit('year');
var x_axis = new Rickshaw.Graph.Axis.ExtendedTime(
{
graph: graph,
tickSpacing: 60*60*24*365*13, // 13 years
timeUnit: timeUnit
} );
to have ticks spaced out evenly at every 13 years.
For Value-based Axes
For value-based Axes, you would need to extend the render() function to include a facility that "manually" sets the ticks for the axis. I did it like this:
this.render = function() {
if (this.graph.height !== this._renderHeight) this.setSize({ auto: true });
var axis = d3.svg.axis().scale(this.graph.y).orient(this.orientation);
if (this.tickSpacing) {
var tickValues = [];
var min = Math.ceil(axis.scale().domain()[0]/this.tickSpacing);
var max = Math.floor(axis.scale().domain()[1]/this.tickSpacing);
for (i = min * this.tickSpacing; i < max; i += 1) {
console.log(i);
tickValues.push(i * this.tickSpacing);
}
axis.tickValues(tickValues);
}
axis.tickFormat( args.tickFormat || function(y) { return y } );
if (this.orientation == 'left') {
var berth = this.height * berthRate;
var transform = 'translate(' + this.width + ', ' + berth + ')';
}
if (this.element) {
this.vis.selectAll('*').remove();
}
this.vis
.append("svg:g")
.attr("class", ["y_ticks", this.ticksTreatment].join(" "))
.attr("transform", transform)
.call(axis.ticks(this.ticks).tickSubdivide(0).tickSize(this.tickSize));
var gridSize = (this.orientation == 'right' ? 1 : -1) * this.graph.width;
this.graph.vis
.append("svg:g")
.attr("class", "y_grid")
.call(axis.ticks(this.ticks).tickSubdivide(0).tickSize(gridSize));
this._renderHeight = this.graph.height;
};
The important part here are the statements in the if (this.tickSpacing) clause. They compute ticks given by the tickSpacing variable in the config array, and assign them to the axis in the axis.tickValues(tickValues) statement. Note that this.tickValues is assigned in the this.tickSpacing = args.tickSpacing statement in the initialize() function, not stated above.
Try it yourself
Have a look at this jsfiddle, where the complete code is available. This will certainly give you some pointers. If you want, you can create your own jsfiddle with your values and tell me if you need anything else.

tickSize is the size of the ticks in pixels. Not what you want to be setting to a huge number.
Set ticks to the number of ticks you want on the graph and Rickshaw (actually d3) will do some magic to give you pretty values of ticks that generate about that number of ticks on the graph.
If you want further control you're going to have to dig into d3, where you will be able to explicitly set the tick values using axis.tickValues(). I'd probably copy the existing Rickshaw.Graph.Axis.Y code and create my own Y axis class that includes access to tickValues or the ability to use my own scale. It's a little unclean in that Rickshaw creates the Y scale in the graph.render() function, so you can't easily override the Y scale, but the Y scale Rickshaw creates does have the range set from the graph data, which is information you will want when creating your own tick values.