I'm currently using the whole idea of
var myQuantizeFunction = d3.scale.quantize()
.domain(minMaxFromData) // the minmax using d3.extent
.range(['bla-1', 'bla-2', 'bla-3', 'bla-4', 'bla-5']);
So this works fine when you want to generate a legend across your min-max. The issue is, I have some data which comes back as 0.
Here is an example legend for context :
As you can see, it's first or lowest value from the range is 0 - 4.7, what I want to really do is have 0 (ie none) as it's own legend item and have everything above ie 1 - 33 in this case as the other ranges.
I want to be able to specify that the first range is 0 and then the domain is split equally between values > 0.
Is there a d3 way of doing this? I'm sure someone else must have had this same problem before, I can't seem to find it but I may not be using the right search terms.
From the documentation:
quantize.domain([numbers])
If numbers is specified, sets the scale's input domain to the
specified two-element array of numbers. If the array contains more
than two numbers, only the first and last number are used. If the
elements in the given array are not numbers, they will be coerced to
numbers; this coercion happens similarly when the scale is called.
Thus, a quantize scale can be used to encode any type that can be
converted to numbers. If numbers is not specified, returns the scale's
current input domain.
As the name suggests d3 is 'data driven' so ignoring parts of your data set is not part of its ethos.
You need to write your own function to generate the [numbers] array.
Try:
data = [0,0,2,1,4,6,7,8,4,3,0,0];
min = undefined;
data.forEach(function (v) {
if (v > 0) {
if (typeof(min) === 'undefined') {
min = v;
} else if (v < min) {
min = v;
}
}
})
var myQuantizeFunction = d3.scale.quantize()
.domain([min, d3.max(data)])
.range(['bla-1', 'bla-2', 'bla-3', 'bla-4', 'bla-5']);
I improved the latest solution to use d3.min() and added code to the test the quantize function. Also I added a small function to colorize the output.
Everything done in the d3 datadriven way
data = [0,0,2,1,4,6,7,8,4,3,0,0];
range = ['bla-1', 'bla-2', 'bla-3', 'bla-4', 'bla-5'];
//strip the first element
reducedRange = range.slice();
reducedRange.shift();
var myQuantizeFunction =
d3.scale.quantize()
.domain([d3.min(data), d3.max(data)])
.range(reducedRange);
var filterQuantize = function(d){
if(d==0){
return range[0];
}else{
return myQuantizeFunction(d);
}
}
var colorize = d3.scale.category10();
// To test this we will put all the data in paragraphs
d3.select('body').selectAll('p').data(data) .enter()
.append('p')
.text(function(d){return d+':'+filterQuantize(d);})
.style('color',function(d){return colorize(d)});
View this code runing
Hope this helps, good luck!
Update: I stripped zero out of the scale to treat it as a special case as you pointed in the comment.
Related
If the value of f5 cell in a Google Sheet is 1.1000 (a number formatted to 4 decimal places) and the value of f6 is = f5 * 1.073, how can I ensure I get the same result multiplying those values in Javascript, eg:
var original_value = 1.1000;
var derivative_value = original_value * 1.073;
Specifically, my question is - will the result of the Javascript multiplication (derivative_value) be the same as the result of the Google formula (f6)? And if not, how can I make it so that it is?
Context / What I've Tried
For context, this question is part of a larger question I am trying to resolve for which I have set up this JSFiddle.
The JSFiddle has an input for the original_value and an input for the multiplier.
It outputs the result to four decimal places and adds trailing zeros where required (this is the required format for the result).
It is an attempt to check that the Javascript code I am writing will produce the same result as the Google Sheet formula.
[ The JSFiddle has been updated to also log decimal.js results to the console for comparison ]
Edit
There was a suggestion to use decimal.js but I'm not sure how it would be applied - something like the following?
var original_value = new Decimal(1.1000);
// some different multipliers for testing
var multiplier_a = new Decimal(1.073);
var multiplier_b = new Decimal(1.1);
// some different results for testing
var derivative_value_a = original_value.times(multiplier_a).toString();
var derivative_value_b = original_value.times(multiplier_b).toString();
console.log(derivative_value_a); // 1.1803
console.log(derivative_value_b); // 1.21
Is that any more accurate than plain Javascript original_value * multiplier? More importantly for this question, will it always simulate the same result that a Google Sheet formula produces?
JavaScript is using so called double precision float format (64 bit)- https://tc39.github.io/ecma262/#sec-terms-and-definitions-number-value
Google Sheets seem to use the same format, you can test it by =f6*1E13 - round(f6*1E13) to see that f6 is not STORED as a fixed number format, only FORMATTED
see Number.toFixed how to FORMAT numbers in Javascript
to generate some test data:
[...Array(10)].forEach(() => {
const f5 = 1.1
const x = Math.random() / 100
const f6 = f5 * x
console.log(x, f6.toFixed(4))
})
and compare in Google Sheet:
https://docs.google.com/spreadsheets/d/1jKBwzM41nwIEyatLUHEUwteK8ImJg334hzJ8nKkUZ5M/view
=> all rounded numbers are equal.
P.S.: you need to copy the console output, paste into the Sheet, use the menu item Data > Split text into columns... > Space, then multiply by 1.1 in 3rd column and finally format all numbers
After revisiting this I have updated the jsFiddle.
The main components of what I believe are a satisfactory solution are:
Convert both original_value and multiplier to decimal.js objects.
Do the multiplication using the decimal.js times method.
Do the rounding using the decimal.js toDecimalPlaces method.
Use the argument values (4,7) to define 4 decimal places with ROUND_HALF_CEIL rounding, equivalent to Math.round (reference)
For example:
var my_decimal_js_value = new Decimal(original_value).times(new Decimal(multiplier)).toDecimalPlaces(4, 7);
In order to add any necessary trailing zeros to the result, I use:
function trailingZeros(my_decimal_js_value) {
var result = my_decimal_js_value;
// add zeros if required:
var split_result = result.toString().split(".");
// if there are decimals present
if (split_result[1] != undefined) {
// declare trailing_zeros;
var trailing_zeros;
// get the amount of decimal numbers
decimals_present = split_result[1].length;
// if one decimal number, add three trailing zeros
if (decimals_present === 1) {
trailing_zeros = "000";
result += trailing_zeros;
}
// if two decimal numbers, add two trailing zeros
else if (decimals_present === 2) {
trailing_zeros = "00";
result += trailing_zeros;
}
// if three decimal numbers, add one trailing zero
else if (decimals_present === 3) {
trailing_zeros = "0";
result += trailing_zeros;
}
// if four decimal numbers, just convert result to string
else if (decimals_present === 4) {
result = result.toString();
}
}
// if there are no decimals present, add a decimal place and four zeros
else if (split_result[1] === undefined) {
trailing_zeros = ".0000";
result += trailing_zeros;
}
return result;
}
I am still not absolutely certain that this mimics the Google Sheet multiplication formula, however using decimal.js, or another dedicated decimal library, seems to be the preferred method over plain JavaScript (to avoid possible rounding errors), based on posts such as these:
http://www.jacklmoore.com/notes/rounding-in-javascript
Is floating point math broken?
https://spin.atomicobject.com/2016/01/04/javascript-math-precision-decimals
I have a product to add into the store, where there are two input fields priceBase and priceFinal. First is without TAX, second is with TAX.
While using this javascript function:
jQuery(function($){
var priceBase = $('input[name="mprices[basePrice][]"]', '#productPriceBody');
var priceFinal = $('input[name="mprices[salesPrice][]"]', '#productPriceBody');
var priceDiff = priceFinal.val() - priceBase.val();
var priceTax = priceDiff / priceBase.val();
alert(priceFinal.val()); // 1.40004
alert(priceBase.val()); // 1.16667
alert(priceDiff); // 0.23333399999999993
alert(priceTax); // 0.19999999999999993
});
How I suppose to round a priceTax value from 0.19999999999999993 to 0.20 ? Like normal math calculation you know, if it's 4 and below, it rounds to lower, if it's 5 it rounds to higher number.
Thanks for suggestions in advance.
You seem to want
alert(priceDiff.toFixed(2));
But you should parse the values before you do maths. It works here because you're lucky :
"33"-"12" => "21"
"33"+"12" => "3312"
So to avoid bugs in the future (when you use + instead of - for example) I'd suggest to always parse the field values :
var priceDiff = parseFloat(priceFinal.val()) - parseFloat(priceBase.val());
I have a Google Chart with a continuous date-time X axis. My data comes in short bursts, with long delays between the bursts. I'd like to make the Chart have a non-continuous X axis, but still have the auto-generated timestamps during the samples. Is that possible?
Basically, say I have 3 samples, each which have 300 datapoints, recorded across 10 second intervals, but with hour gaps between them. I'd like to have my chart show the 30 seconds of data at a zoom level where it can be distinguished. Am I stuck?
Edit: Per #jmac's suggestion, here is an example of what the data looks like:
1360096658270, 10.228335
1360096658274, 10.308437
1360096658277, 10.294770
[...]
1360096673968, 9.014943
1360096673969, 8.971434
1360096673970, 9.041739
1360096673971, 9.097484
^^-- 15 seconds
<--- (~10 days)
1360989176509, 9.856928
1360989176513, 9.852907
1360989176517, 9.861740
1360989176523, 9.820416
1360989176527, 9.871401
Method 1: Multiple Charts
This is probably the simplest in concept (though still a hassle).
Summary:
Split data in to groups (eliminate the gaps)
Create a separate chart for each group
Eliminate the vAxis labels for every chart past the first
Create a consistent vAxis min/max value
Use CSS to line the charts up side to side
Details:
If you have a static data set, you can just split it by hand. If it isn't static, then you have to write some javascript to split up your data. I can't really help you here since I don't know how your data works.
As far as setting up the charts, I'll leave that up to you. I don't know how you want them formatted, so again I can't really help you with the current info.
To create a consistent axis value for all charts, you need to use some basic math in a javascript function to assign the same numbers to each vAxis max/min value. Here is a sample:
// Take the Max/Min of all data values in all graphs
var totalMax = 345;
var totalMin = -123;
// Figure out the largest number (positive or negative)
var biggestNumber = Math.max(Math.abs(totalMax),Math.abs(totalMin));
// Round to an exponent of 10 appropriate for the biggest number
var roundingExp = Math.floor(Math.log(biggestNumber) / Math.LN10);
var roundingDec = Math.pow(10,roundingExp);
// Round your max and min to the nearest exponent of 10
var newMax = Math.ceil(totalMax/roundingDec)*roundingDec;
var newMin = Math.floor(totalMin/roundingDec)*roundingDec;
// Determine the range of your values
var range = newMax - newMin;
// Define the number of gridlines (default 5)
var gridlines = 5;
// Determine an appropriate gap between gridlines
var interval = range / (gridlines - 1);
// Round that interval up to the exponent of 10
var newInterval = Math.ceil(interval/roundingDec)*roundingDec;
// Re-round your max and min to the new interval
var finalMax = Math.ceil(totalMax/newInterval)*newInterval;
var finalMin = Math.floor(totalMin/newInterval)*newInterval;
Method 2: Multiple Series
As long as the people viewing your data understand they are different sets, then there's no reason the axis needs to say the exact date/time as long as they can easily figure that out elsewhere.
Summary:
Separate your data in to different series for each 'sequence'
Artificially shorten the gaps between sequences (if they are 15 seconds each, then have a 5 second gap between series, or just start every 15 seconds)
Format each different series with a name labeling when the run started/ended
Details:
Again, you will have to split your data manually or create javascript to do it, but what you want to do is to move each set of numbers in to its own column, like so:
1360096658270, 10.228335, null
1360096658274, 10.308437, null
1360096658277, 10.294770, null
[...]
1360096673968, 9.014943, null
1360096673969, 8.971434, null
1360096673970, 9.041739, null
1360096673971, 9.097484, null
^^-- 15 seconds
<--- (~10 days)
1360989176509, null, 9.856928
1360989176513, null, 9.852907
1360989176517, null, 9.861740
1360989176523, null, 9.820416
1360989176527, null, 9.871401
This will make each series be a different color (and have a different label in the legend/on mouseover), so you can see the difference between runs, but also get a nice tooltip saying "This data was gathered from X to Y" so that if the time the data was taken is important, it's still in there (albeit not on the X axis).
These are the easiest ways.
Method 3: Manually Editing the X-Axis Labels
The third way is the most flexible but also takes the most work. You can create a custom javascript function to manipulate the X-axis labels in SVG. More details on this here by #jeffery_the_wind:
/*
*
* The following 2 functions are a little hacky, they have to be done after calling the "draw" function
* The bubble chart originally displays only numbers along the x and y axes instead of customer or product names
* These 2 functions replace those numbers with the words for the customers and products
*
*/
for ( var i = -2; i < products.length + 1; i ++ ){
$('#customer_product_grid svg text[text-anchor="start"]:contains("'+i+'")').text(function(j,t){
if (t == i){
if (i >= products.length || i < 0){
return " ";
}
return products[i];
}
});
}
for ( var i = -2; i <= customers.length + 3; i ++ ){
$('#customer_product_grid svg text[text-anchor="end"]:contains("'+i+'")').text(function(j,t){
if (i >= customers.length + 1 || i <= 0){
return " ";
}else if (t == i){
return customers[i-1];
}
});
}
Google's documentation on customizing axes describes how to do what you're asking. You can change the type of your column to a string and populate with formatted Date strings.
While playing around with random numbers in JavaScript I discovered a surprising bug, presumably in the V8 JavaScript engine in Google Chrome. Consider:
// Generate a random number [1,5].
var rand5 = function() {
return parseInt(Math.random() * 5) + 1;
};
// Return a sample distribution over MAX times.
var testRand5 = function(dist, max) {
if (!dist) { dist = {}; }
if (!max) { max = 5000000; }
for (var i=0; i<max; i++) {
var r = rand5();
dist[r] = (dist[r] || 0) + 1;
}
return dist;
};
Now when I run testRand5() I get the following results (of course, differing slightly with each run, you might need to set "max" to a higher value to reveal the bug):
var d = testRand5();
d = {
1: 1002797,
2: 998803,
3: 999541,
4: 1000851,
5: 998007,
10: 1 // XXX: Math.random() returned 4.5?!
}
Interestingly, I see comparable results in node.js, leading me to believe it's not specific to Chrome. Sometimes there are different or multiple mystery values (7, 9, etc).
Can anyone explain why I might be getting the results I see? I'm guessing it has something to do with using parseInt (instead of Math.floor()) but I'm still not sure why it could happen.
The edge case occurs when you happen to generate a very small number, expressed with an exponent, like this for example 9.546056389808655e-8.
Combined with parseInt, which interprets the argument as a string, hell breaks loose. And as suggested before me, it can be solved using Math.floor.
Try it yourself with this piece of code:
var test = 9.546056389808655e-8;
console.log(test); // prints 9.546056389808655e-8
console.log(parseInt(test)); // prints 9 - oh noes!
console.log(Math.floor(test)) // prints 0 - this is better
Of course, it's a parseInt() gotcha. It converts its argument to a string first, and that can force scientific notation which will cause parseInt to do something like this:
var x = 0.000000004;
(x).toString(); // => "4e-9"
parseInt(x); // => 4
Silly me...
I would suggest changing your random number function to this:
var rand5 = function() {
return(Math.floor(Math.random() * 5) + 1);
};
This will reliably generate an integer value between 1 and 5 inclusive.
You can see your test function in action here: http://jsfiddle.net/jfriend00/FCzjF/.
In this case, parseInt isn't the best choice because it's going to convert your float to a string which can be a number of different formats (including scientific notation) and then try to parse an integer out of it. Much better to just operate on the float directly with Math.floor().
I'd like to create a random boolean in JavaScript, but I want to take the previous value into account. If the previous value was true, I want it to be more likely for the next value to be true. At the moment I've got this (this is in the context of a closure - goUp and lastGoUp are locals to the containing scope):
function setGoUp() {
goUp = getRandomBoolean();
if(lastGoUp) {
goUp = getRandomBoolean() || goUp;
}
else {
goUp = getRandomBoolean() && goUp;
}
lastGoUp = goUp;
}
So, the algorithm goes:
Get a random boolean
If the random boolean from the previous call was True:
a) get another random boolean, and or these two together
b) else get another random boolean and and these together.
I'm sure this algorithm could be simplified. I wondered about doing:
if(lastGoUp && goUp) {
goUp = goUp * (getRandomBoolean() || goUp);
}
but that seems really dirty.
There's also a problem with this algorithm which means that I can only double the chance of getting the same boolean again - I can't tweak it easily. Any ideas?
You should define the distribution you want, but maybe you are looking for the following?
if (lastGoUp) {
goUp = Math.random() < 0.8;
} else {
goUp = Math.random() < 0.2;
}
Instead of getting a random boolean, get a random number, say between 0 and 99. Keep a threshold value instead of the last number, and adjust the threshold according to the result:
var threshold = 50;
function setGoUp() {
goUp = getRandomNumber() < threshold;
threshold += goUp ? -10 : 10;
}
This would keep a running tab, so if you get consecutive results that are the same, the probability would keep falling for that result.
If you only want to consider the last result, you would instead set the threshold to a specific value:
threshold = goUp ? 40 : 60;
If you only want the probability of the next event to depend on the current value, and not the history of values up til now, what you want is called a Markov process. Often these are implemented with a 2D table of probabilities that you look up (prob of each next outcome given current one), but for a simple bool-valued event, an if statement is sufficient (see meriton's answer; note that it corresponds to a table of probabilities [0.8 0.2; 0.2 0.8]).
If you want something that gets more likely, say, the more successes you get in a row, then you need to devise a sequence of probabilities for success that perhaps approaches, but does not exceed, 1. There are any number of formulas which can do this, depending on how strong you want the bias to become and how quickly you want it to get there.
I would just make the probability of getting value true be an explicit float variable p. Then I could tweak it easily, by increasing p in some way if I got true last time or by doing nothing with it if I got 'false'.
Can replace Math.random for a better randomizer.
var setGoUp = (function(){
var last;
return function(){
// if last 66% chance for true else 50% chance of true.
return !!(last ? Math.random()*3 : Math.random()*2);
}
}());
!! converts anything to a boolean, 0 = false.