Plotting large dataset using JQPlot - javascript

I have an application that needs to plot at least over 5k data points and the size can nearly be limitless.
The plot goes:
X-Axis -> DateTime
Y-Axis -> Temperature
This is my current plot
var plot1 = $.jqplot('chartdiv', dataArray, {
title: 'Default Date Axis',
axes: { xaxis: { renderer: $.jqplot.DateAxisRenderer } },
series: [{ lineWidth: 4, markerOptions: { style: 'square' } }],
numberTicks: 10
});
with data array containing at least 5k points in the ['Date', 'Temperature'] format.
The problem with this is that it'x extremely inefficient and makes the browser freeze up. I don't need it to literally put a label on every datapoint, maybe a few. Can anyone give me tips on how to optimize this?

You can get better performance by changing your marker types and removing shadows. I've been using these options in my plots with pretty good success:
seriesDefaults: {
lineWidth: 1, shadow: false,
rendererOptions: { smooth: false },
markerOptions: { show: true, shadow: false, size: 2 }
}
Here's an example which adds 50k points to a jqplot: http://jsfiddle.net/xf8d36kc/
We have a project where we are consistently plotting 20-30k points fairly quickly. Once we start getting towards around 100k points it's still quick to display but zooming and interacting with the plot starts to have lag times of 1-3 seconds but still doesn't freeze up the browser. We experience more delay in bringing the data in from our database then actual draw time for jqplot.
I haven't looked much in to plotting sets much bigger than 100k, but after that you may want to start grouping the data serverside to limit the number of points and then expand them as you zoom in for more detail.

If working with Primefaces 5, do this:
<p:chart id="lineChart" type="line" model="#{managedBean.lineModel}" />
<h:outputScript>
function chartExtender() {
// this = chart widget instance
// this.cfg = options
this.cfg.seriesDefaults = {
lineWidth: 2, shadow: false,
rendererOptions: { varyBarColor: true, smooth: false },
markerOptions: { show: true, shadow: false }
};
}
</h:outputScript>
On Managedbean:
LineChartModel lineModel = new LineChartModel();
//...
lineModel.setExtender("chartExtender");
//...
Greatly improved the performance for me.
Others options here.

Related

How to make one line smooth and the other NOT on a chart?

So I'm looking for a way to make one of the lines on my chart NOT smooth and the other smooth. There is a master option to make every line on the chart smooth but not a particular line that I know of.
Here is my chart of what it looks like
PS. Im using JQ Plots http://www.jqplot.com/index.php
Here is my JS code.
//STOCK HIGH VS LOW
var line1 = [['2017-11-27',175.0800],['2017-11-28',174.8700],['2017-11-29',172.9200],['2017-11-30',172.1400],['2017-12-01',171.6700],['2017-12-04',172.6200],['2017-12-05',171.5200],['2017-12-06',170.2047],['2017-12-07',170.4400],['2017-12-08',171.0000],['2017-12-11',172.8900],['2017-12-12',172.3900],['2017-12-13',173.5400],['2017-12-14',173.1300],['2017-12-15',174.1700],['2017-12-18',177.2000],['2017-12-19',175.3900],['2017-12-20',175.4200],['2017-12-21',176.0200],['2017-12-22',175.4240],['2017-12-26',171.4700],['2017-12-27',170.7800],['2017-12-28',171.8500],['2017-12-29',170.5900],['2018-01-02',172.3000],['2018-01-03',174.5500],['2018-01-04',173.4700],['2018-01-05',175.3700],['2018-01-08',175.6100],['2018-01-09',175.0600],['2018-01-10',174.3000],['2018-01-11',175.4900],['2018-01-12',177.3600],['2018-01-16',179.3900],['2018-01-17',179.2500],['2018-01-18',180.1000],['2018-01-19',179.5800],['2018-01-22',177.7800],['2018-01-23',179.4400],['2018-01-24',177.3000],['2018-01-25',174.9500],['2018-01-26',172.0000],['2018-01-29',170.1600],['2018-01-30',167.3700],['2018-01-31',168.4417],['2018-02-01',168.6200],['2018-02-02',166.8000],['2018-02-05',163.8800],['2018-02-06',163.7200],['2018-02-07',163.4000],['2018-02-08',161.0000],['2018-02-09',157.8900]];
var line2 = [['2017-11-27',173.3400],['2017-11-28',171.8600],['2017-11-29',167.1600],['2017-11-30',168.4400],['2017-12-01',168.5000],['2017-12-04',169.6300],['2017-12-05',168.4000],['2017-12-06',166.4600],['2017-12-07',168.9100],['2017-12-08',168.8200],['2017-12-11',168.7900],['2017-12-12',171.4610],['2017-12-13',172.0000],['2017-12-14',171.6500],['2017-12-15',172.4600],['2017-12-18',174.8600],['2017-12-19',174.0900],['2017-12-20',173.2500],['2017-12-21',174.1000],['2017-12-22',174.5000],['2017-12-26',169.6790],['2017-12-27',169.7100],['2017-12-28',170.4800],['2017-12-29',169.2200],['2018-01-02',169.2600],['2018-01-03',171.9600],['2018-01-04',172.0800],['2018-01-05',173.0500],['2018-01-08',173.9300],['2018-01-09',173.4100],['2018-01-10',173.0000],['2018-01-11',174.4900],['2018-01-12',175.6500],['2018-01-16',176.1400],['2018-01-17',175.0700],['2018-01-18',178.2500],['2018-01-19',177.4100],['2018-01-22',176.6016],['2018-01-23',176.8200],['2018-01-24',173.2000],['2018-01-25',170.5300],['2018-01-26',170.0600],['2018-01-29',167.0700],['2018-01-30',164.7000],['2018-01-31',166.5000],['2018-02-01',166.7600],['2018-02-02',160.1000],['2018-02-05',156.0000],['2018-02-06',154.0000],['2018-02-07',159.0685],['2018-02-08',155.0300],['2018-02-09',150.2400]];
var plot2 = $.jqplot('chart1', [line1, line2], {
title: 'AAPL High vs Low',
seriesDefaults: {
rendererOptions: {
//////
// Turn on line smoothing. By default, a constrained cubic spline
// interpolation algorithm is used which will not overshoot or
// undershoot any data points.
//////
smooth: true
}
},
legend: { show: true },
axes: {
xaxis: {
renderer: $.jqplot.DateAxisRenderer,
tickOptions: { formatString: '%b %#d, %#I %p' },
min: 'November 20 2017',
tickInterval: '1 month'
}
},
highlighter: {
show: true,
sizeAdjust: 7.5
},
canvasOverlay: {
show: true,
objects: [
{horizontalLine: {
name: 'average',
y: 171.21, //**AVERAGE_FLOAT_VALUE**
lineWidth: 2,
color: '#FF5555',
shadow: false
}}
]
},
series: [{ lineWidth: 4,
markerOptions: { style: 'square' }
}],
series: [
{ label: 'High' },
{ label: 'Low' }
],
});
If it cant be done in JQ Plots, do you guys know of any other charts that can do this.
Pretty much Im looking for a chart to look like this in the end.
Replying kinda late but..
Yes you can do it in jqplot as well.
Instead of using a seriesDefaults option, go directly and add it in each series you want:
series: [
{
// custom line
lineWidth:2,
markerOptions: { style:'diamond' }
},
{
// here is your smooth line
rendererOptions: {
smooth: true,
},
markerOptions: { style:"circle" }
}
]
Disclaimer: I read it on the internet . . . but apparently d3 can do this (https://gist.github.com/thiyagu-p/3925981). Also, if you Google 'd3 moving average and volume' you get some interesting results. YMMV.
I don't know about jqPlot but you could certainly do this with RGraph:
https://www.rgraph.net/demos/svg-line-trading.html
And if you wanted the gray decorative bar at the back you could use another Bar chart to get this effect before drawing the Line charts and give it some dummy data of:
[1,0,1,0,1,0,1,0,1,0,1,0,1,0,1]

JQPLOT bar graph range

I have created a dynamic bar graph using jqplot, the problem is that some values are quite big i.e. 10000000 and others are pretty small i.e. 2000 or 10000. Because of this range issue, i am only able to see the bigger valued graphs while the rest dont even appear up or appear as just a line on the axis.
Could someone please let me know how this range issue could be solved?
Following my comment, and your need to achieve a solution which includes jqplot bar chart, i'll suggest illustrating the massive difference of values by adding a second line graph to the plot.
$(document).ready(function(){
var data = [10000000, 5000000,10000,2000];
var options= {
title: 'Bar and Line Chart',
series:[{
renderer: $.jqplot.BarRenderer,
rendererOptions: {
barMargin: 5,
fillToZero: true
},
pointLabels: {
show: true,
seriesLabelIndex:1,
hideZeros:false
}
}],
axes: {
xaxis: {
tickRenderer: $.jqplot.CanvasAxisTickRenderer,
renderer: $.jqplot.CategoryAxisRenderer,
},
yaxis: {
labelRenderer: $.jqplot.CanvasAxisLabelRenderer,
padMin: 0,
pad: 1.1,
label: ' logarithmic scale',
rendererOptions: { forceTickAt0: true}
}
},
};
$.jqplot('graph', [data,data], options );
});
Here is a working solution in jsfiddle

Highstock area chart renders a gap, but there is data

My highstock chart isn't rendering properly. It shows some gap when my 2nd series start in the area chart. Even when zoomed in it's not showing properly. The timestamps are in UTC and have a value.
To clarify it a bit here some screenshots:
Normally not zoomed in:
Zoomed in a bit, and showing mouseover. Has only 1 series:
Zoomed in more, showing bigger gap, but has only 1 series:
Zoomed in more, showing bigger gap, but has only 1 series on same date:
I have the following options enabled on the graph:
chart: {
zoomType: 'x',
type: 'area'
},
plotOptions: {
area: {
stacking: 'normal'
},
series: {
turboThreshold: 0,
dataGrouping: {
approximation: "high",
smoothed: true,
groupPixelWidth: 10,
units: [['day', [1]], ['week', [1]], ['month', [1]]]
}
}
}
Any suggestions how to resolve this?
Thanks!
Seems that disabling
plotOptions.series.dataGrouping.smoothed
does the trick.
So then a bug-report:
When
plotOptions.series.dataGrouping.smoothed = true
the graph renders gaps, where I would expect a line.

How to stop Flot Charts from re-ordering my data?

I am using a line chart. I feed the data the following:
var scheduled = [[51,1700],[52, 1750],[1,1600],[2,1675]];
var actual = [[51,1320],[52, 1550],[1,1575],[2,1600]];
In the above the first number of each set is the week of the year and I am trying to show the last 4 months of data.
However, when the chart is drawn Flot charts re-sorts the data by the first value (lowest to highest) which creates all kinds of issues. Instead of 4 columns in the series there are now 52, and the lines are quite out of whack.
I don't see anything in the documentation that says this is supposed to happen, nor do I see anything that says I can prevent it. However, for the data to be meaningful, the data must not be re-ordered.
Is there a setting I'm unaware of that can stop this behavior?
Edit : Adding plot code
var plot = $.plot('#scheduled-actual-flot-line', [
{
label: 'Scheduled Hours',
data: scheduled,
lines: { show: true, lineWidth: 2, fill: true, fillColor: { colors: [{ opacity: 0.5 }, { opacity: 0.5 }] } },
points: { show: true, radius: 4 }
},
{
label: 'Actual Hours',
data: actual,
lines: { show: true, lineWidth: 2, fill: true, fillColor: { colors: [{ opacity: 0.5 }, { opacity: 0.5 }] } },
points: { show: true, radius: 4 }
}],
{
series: {
lines: { show: true },
points: { show: true },
shadowSize: 0 // Drawing is faster without shadows
},
colors: ['#afd2f0', '#177bbb'],
legend: {
show: true,
position: 'nw',
margin: [15, 0]
},
grid: {
borderWidth: 0,
hoverable: true,
clickable: true
},
yaxis: { ticks: 4, tickColor: '#eeeeee' },
xaxis: { ticks: 12, tickColor: '#ffffff' }
}
);
Flot takes the x values as numbers and displays / sorts them accordingly. If you don't want that, you can use the category mode (see this example and this fiddle with your data).
xaxis: {
//ticks: 12,
tickColor: '#ffffff',
mode: 'categories'
}
PS: 12 ticks are not possibly with your data, as there are only 4 datapoints defined.
That flot reads all data as numbers by default is described here in the documentation.
Flotr examples use a for loop to create random data, so the first index will always be sequential.
[[51,1700],[52, 1750],[1,1600],[2,1675]];
Your arrays show that flotr must be doing a sort on the array before painting the data sets as lines, bar-graphs or whatever.
I can only suggest you create a timestamp from the months and there's a time setting you can in flotr settings to format the dates as you want.
The other way is replace your anomalous data (months) with sequential indices:
var arr = [[51,1700],[52, 1750],[1,1600],[2,1675]];
for(var i=0; i<arr.length; i++) arr[1][0] = i;
Flot is doing exactly what it should do for a line chart (or any type of x-y graph). It's showing the last two points of your dataset on the left because 1 and 2 are indeed less than 51 and 52. I'm guessing that you're trying to show data that crosses a year boundary. You need to make the first two weeks of the second year later than the last two of the first. You could use actual dates instead of week numbers, in which case Flot would handle it fine. That would also give you more flexibility in labeling the x-axis. But as a quick fix, just add 52 to the second year's data, e.g.:
var scheduled = [[51,1700],[52, 1750],[53,1600],[54,1675]];
var actual = [[51,1320],[52, 1550],[53,1575],[54,1600]];

Horizontal Bar Chart Ignores tickInterval, min and max Options

I have been using jqPlot for a project I am in and so far I am quite satisfied with it. Is very complete. However, the feature I am working on right now requires an horizontal bar chart. My issue is that my chart currently shows xaxis ticks as decimal numbers increasing by 0.5. I don't need decimal values. I need integers increasing by 1. My config is a little something like this:
var chartConfig = {
seriesDefaults: {
showMarker: false,
shadowAngle: 135,
pointLabels: {
show: true,
labelsFromSeries: true,
seriesLabelIndex: 2,
edgeTolerance: -50
},
renderer: $.jqplot.BarRenderer,
rendererOptions: {
barWidth: 25,
barPadding: 0,
barDirection: 'horizontal'
}
},
axes: {
xaxis: {
min: 0,
max: 4,
tickInterval: 1
},
yaxis: {
showTicks: true,
renderer: $.jqplot.CategoryAxisRenderer,
tickRenderer: $.jqplot.CanvasAxisTickRenderer,
tickOptions: {
angle: -50,
labelPosition: 'middle'
}
}
}
};
However, looks like jqPlot is completely ignoring the xaxis tickInterval, min and max option values. My output chart looks as follows:
I need to know what I am doing bad here. My config looks to be okay, but I can't get jqPlot to use the values I pass for the X axis. Any help on this matter will be very much appreciated.
Thanks.
I have been having problem with the tickInterval on horizontal bars also. I found that you have to specify both the min and max along with the tick interval for the axis you need it for.
This you have done but I wanted to state this somewhere as I spent a long time trying to find why tick interval wasn't working and this post kept coming up in my searches so hopefully it will help others.
For your issue, is possibly something to do with your data sets.
If this is still causing you problems maybe post the full script with your values as the code is only partial atm.
Thanks
Richard

Categories