Table of Contents

(End of Page)

Introduction

Back to top
There are many ways to make scatter plots, graphs, bar charts, pie charts, etc in javascript. Google developers have a wonderful set of free javascript sources that you can use called
Google Charts that are quite powerful, and SitePoint has a whole page of javascript libraries for this purpose.

So why develop yet another? Partly because I can, that is, to learn now. But also because any most of the libraries have to be downloaded, so you need to always have the internet, and good access at that. Because any library that is supported by companies like Google grow to be large and complex and easy to use unless you want to do something they haven't intended. Because the documentation is not all that great (like much documentation, it's written by the developers for the developers, and not for the users as much). Because for science purposes (which is what I am pursuing), we don't need fancy bar charts, etc, we just need histograms, scatter plots, and functions (more or less) whereas other libraries are of course comprehensive, and very useful for other things (like business and administration). And so on. Anyway, the UMD Javascript plotting libraries are pretty easy to use, sort of modeled on the Google Charts, lightweight, and downloadable so you don't need to access it over the internet.

What is usually the case in science is that we need histograms, scatterplots, and functions, and this is what this package provides. No bar charts, pie charts, etc, just basic histograms and scatter plots (functions are made from scatterplot functions).


Version

Back to top
You can get the version by doing the following in your javascript code:
	var version = umd_show_version();
	alert(version);
or
	var version = umd_show_version();
	console.log("Version ="+version);

Access and Usage

Back to top
The code is available for download
here. There are 3 parts to using this code:
  1. In your html, preferably inside the <head> tag, include the following javascript source:
    	<script src="/location/plots.js"></script>
    
    where "location" is of course dependent on where you put the code.
  2. All plots, whether histograms,scatter plots, or functions (or whatever is implemented) are drawn in an HTML5 canvas (<canvas>). The code also will use an HTML div element (<div>) to allow you to do things like click on a histogram bin and get the bin value and contents, etc (see below). So, also inside your html, you need to create a canvas and div tag, each with an id. For example:
    <canvas id="mycan"></canvas>
    <div id="mydiv"></div>
    
    The key thing here is the canvas id and div id string. Keep track of it, you will need it to instantiate the plot you want to make.
  3. Instantiate a histogram, scatter plot, or function inside your javascript source, and refer to the canvas and div id string explicitly (in the above example, "mycan" and "mydiv"). See below for details.

Javascript

The first step in making plots is of course to come up with the data. You do this by: Once you have the data, there's a 3 step process you need to be aware of in order to make and display your plots:
  1. Instantiate a new object (histogram, scatter plot, function....):
  2. Set options (see below) using the Javascript variable type called "literal", with the following example:
    	var options = {
    		opt1 : value1,
    		opt2 : value2,
    		opt3 : value3
    	}
    
  3. Draw the plot

So to put it all together, the following javascript is what you might write to make a histogram of some data:

	//
	// make and array of ages, and fill it somehow
	//
	var xdata = [];
	.
	. fill the array...
	.
	//
	// "book" a histogram with the following arguments:
	//		canvas id
	//		histogram title
	//		number of bins
	//		low edge of 1st bin
	//		upper edge of last bin
	//
	// so here, "mycan" matches the canvas id in the <canvas> tag, followed by
	// the title, and the histogram will have 20 bins between 0 and 80
	//
	var h1 = new umd_histogram("mycan","Age Distribution Histogram",20,0,80);
	var hopt = {
		xtitle : "Age",
		ytitle : "Number"
	};
	h1.plot(xdata,hopt);
The utility will make the histogram and display it inside the canvas.

Functions are even easier:

	var fopt = {
		xtitle : "x",
		ytitle : "x*exp(x)"
	};
	umd_histogram("mycan","Function plot",myfun,fopt);
	//
	function myfun(x) {
		return x*Math.exp(x);
	}

Interactions

Back to top

Interactions

As described below in the options section on "active", you can interact with the plot in two ways get the actual data:

Options

Back to top
Option Description Example Default

Options used by both histogram and scatterplots

active Allows you to click to on histogram box to read out data points. Also there's a white box in the upper left corner, click to read out all data points. active:true active:false
active_div HTML <div> element. You have to add this to your HTML after you give the canvas, e.g.
active_precision Set the precision for rounding numbers placed in the active_div 5active_precision: 3
active_exponential Output the numbers placed into active_div in exponential form active_exponential: true active_exponential: false
debugdebugging, with output to the consoledebug:true debug:false
backgroundbackground color"grey" or "#A5A5A5" default
borderhistogram border (drawn in black)border:true border:false
titletitle stringtitle:"My Title" default to instantiation value
title_fonttitle fonttitle_font:"14px arial" title_font:"#12px Courier"
xtitlehorizontal axis title stringxtitle:"Age" none
xtitle_position"left", "center", "right" (case insensitive) xtitle_position: "right" xtitle_position: "center"
widthcanvas width in pixelswidth:500 width:400
heightcanvas height in pixelsheight:500 height:400
show_statsshows stats on upper right sideshow_state:true show_state:true
gridturn grid on/off grid:false grid:true
xgridnumber of grid lines along x-axisxgrid:5 xgrid:5
ygridnumber of grid lines along y-axisygrid:5 ygrid:6
yminvertical axis min for displayingymin:0 minimum bin content
ymaxvertical axis max for displayingymax:20 maximum bin content

Histogram only Options

normalize normalize area under histogram normalize:1 not enabled
hist_color color of histogram rectange for displaying hist_color:"yellow" hist_color:"#4070d0"
r_fract fraction of bin taken up by histogram rectangle r_fract:1 r_fract:0.9
f_labels 1/fraction of all bins that have a label f_label:4 f_label:5

Scatter Plot only Options

xminhorizontal min value for displayingxmin:1 minimum x value from array
xmaxhorizontal max value for displayingxmax:1 maximum x value from array
yminvertical min value for displayingymin:1 minimum y value from array
xtick_unitunit for ticks and labels along x-axis xtick_unit: 0.1(xmin-xmax)/xgrid
xtick_unit_precisionnumber of decimal places to show for labels along x-axis xtick_unit_precision: 5xtick_unit_precision: 3
ytick_unitunit for ticks and labels along y-axis ytick_unit: 0.1(ymin-ymax)/ygrid
ytick_unit_precisionnumber of decimal places to show for labels along y-axis ytick_unit_precision: 5ytick_unit_precision: 3
point_radiusradius of circle used for points point_radius:2point_radius:3
point_colorcolor of circle used for points point_color:"blue"point_color:"#4070d0"
line_colorcolor of line for plot_type="line" line_color:"blue"line_color:"black"
logplot on log scale log:truelog:false
plot_type"points", "line", "both" plot_type:"line"plot_type:"points"

Function only Options

xlow low value of x xlow:0 min x returned by the function
xhigh high value of x xhigh:10 max x returned by the function
numpoints number of different x values to plot numpoints:200 numpoints:100


Overlay

Back to top
Often, you might want to make a histogram and then overlay a plot of data on top of it. This is quite easy to do here.

For both histograms and scatter plots, after you invoke the .plot method, you can then invoke .overlay, and provide new data and options as arguments. For instance, to overlay a function on top of a histogram, you might do something like this:

	.
	.
	.
	var h = new umd_histogram("mydif","Title",nbins,xlow,xhigh);
	var hdata = [];
	var hopts = {
		opt1 : value,
		opt2 : value,
		.
		.
		.
	}
	h.plot(h,hopts);
	//
	// now add data on top
	//
	var ydata = [];
	for (var i=0; i<n; i++) ydata.push(...);		// this fills ydata
	var overlay_options = {
		o_opt1 : value,
		o_opt2 : value
		.
		.
		.
	}
	h.overlay(ydata,overlay_options);
What the code will do is take the ydata values, and assume a 1:1 correspondence with this histogram bins, and plot the data on top. Note that this is ok, because you've specified the number of bins, and the bin edges, so you should be able to figure out what bin number goes with what "ydata[i]" value. The code will overlay 1 point on top of each bin.

For scatter plots, it's much the same thing, only here you give it 2 arrays, x and y, and options, like this:

	.
	.
	.
	var x_overlay = [];
	var y_overlay = [];
	.
	// fill x_overlay and y_overlay
	.
	s.overlay(x_overlay,y_overlay,overlay_options);
For scatter plots, the x_overlay and y_overlay pair do not have to have the same array length or be related to the original scatter plot x/y pair. The code just overlays the plots.

Overlay options

For histogram or scatter plot overlays, you can plot a circle, a square, connect points with lines, or provide your own character to plot. The following options describe how to do this:
OptionDescriptionExample Default
overlay_type0=circle,1=box,2=line,3=customoverlay_type:1 overlay_type:0
overlay_circle_radiusfor type=0 (circle), radius of circle overlay_circle_radius:3 overlay_circle_radius:1
overlay_circle_colorfor type=0 (circle), color of circle overlay_circle_color:"blue" overlay_circle_color:"red"
overlay_box_fractionfor type=1 (rectangle) histograms, box width as fraction of bin overlay_box_fraction:0.5 overlay_box_fraction:0.25
overlay_box_widthfor type=1 (rectangle) scatter plots, box width in pixels overlay_box_fraction:0.5 overlay_box_fraction:0.25
overlay_box_colorfor type=1 (rectangle), color of box overlay_box_color:"green" overlay_box_color:"blue"
overlay_line_widthfor type=2 (line), width of line overlay_line_width:1 overlay_line_width:2
overlay_line_colorfor type=2 (line), color of line overlay_line_color:"yellow" overlay_line_color:"blue"
overlay_characterfor type=3 (custom), the character you want to display overlay_character:"o" overlay_character:"*"
overlay_character_colorfor type=3 (custom), color of character to display overlay_character_color:"blue" overlay_character_color:"green"

Error Bars

Back to top
(Scatter plots only as of May 9, 2017)

We often want to also see the error bars. Do to this, all you have to do is fill an array of values that contain the error bar, and call via:

	sc = new umd_scatterplot(....)
	var xdata = [];
	var ydata = [];
	.
	.
	sc.plot(xdata,ydata,...);

	var errors = [];
	.
	.
	.
	sc.errorbars(errors,options)
The array "errors" should be in phase with the array "xdata" and "ydata" as in the above example, so that the value at "xdata[i]" will be "ydata[i] +- errors[i]". This means that the error bars will be symmetric about the "ydata" value, and the value in "errors" causes a bar to be drawn with equal distance on both above and below the data point.

Error bar options

The only options at this time (5/8/17) are the errorbar width (width of the vertical error bar), the color, and the width of the horizontal top and bottom of the bar:
OptionDescriptionExample Default
errorbar_widthWidth of the vertical error barerrorbar_width:2 errorbar_width:1
errorbar_colorcolor of the error bar errorbar_color:"green" errorbar_color:"black"
errorbar_barwidthwidth of the horizontal upper and lower bars in pixels errorbar_barwidth:0 errorbar_barwidth:5
Back to top
Copywrite Drew Baden, June 21, 2017