Preface: I've gone through all other stackoverflow questions relating to this problem I could find, which are outdated or recommend using a ordinal scale instead. Including this one: D3 Non-Continuous Dates Domain Gives Gaps on X-Axis
which describes my problem as well.
Goal: I am trying to create a candlestick chart with zoom in/zoom out and panning. Currently the chart is working as intended, minus the x-axis (which contains dates/times of each candlestick).
I want the x axis to function like it does with scaleTime(), where it shows years/months, and when you zoom in it shows the date/time, depending on the time interval. I know timeScale() is continuous, but I need the functionality of the zooming in and out.
Problem: Gaps are included in the x-axis since when I set the domain for the x-scale and x-band, I am using the minimum and maximum dates in the dataset (named 'prices').
Code (d3js v7):
Works but x-scale is not dates, but decimal values from -1 to dates.length:
var xScale = d3.scaleTime()
.domain([-1, dates.length])
.rangeRound([0, w]);
Works x-scale is dates, but includes gaps for weekends and holidays:
var xScale = d3.scaleTime()
.domain([xmin, xmax])
.rangeRound([0, w]);
Doesn't work, x-scale is blank, coordinates are messed up:
var xScale = d3.scaleTime()
.domain(prices.map(r => r.t))
.rangeRound([0, w]);
I calculate the x coordinates for my candle bodies for (map) and [xmin,xmax] as:
.attr('x', (d, i) => xScale(d.t) - xBand.bandwidth())
or for domain([-1, dates.length])
.attr('x', (d, i) => xScale(i) - xBand.bandwidth())
xBand is declared as:
let xBand = d3.scaleBand()
.domain(d3.range(-1, dates.length))
.range([0, w])
.padding(0.2);
Another solution I can think of is to somehow convert the demical values from the domain in the first xScale example to function like the scaleTime() scale using tickFormat(), but I'm not sure how I would go about doing that. I have already converted them from decimal to dates using .tickFormat(prices.map(d => new Date(d.t))); but don't know where to go from there.
Questions: Can you pass an array of values as the domain? Or is it just a min and max value, since domain is defined as boundaries within which your data lies?
Am I passing the array mapping correctly? Is this allowed?
Finally, how can I set my domain to the array of dates I have without including gaps on the chart?
Related
I have a scatterplot that uses constraint relaxation to de-conflict the labels for the points that it graphs (Plunker here). The problem is that, when I relax the constraints, this causes collisions between the point labels and the x-axis labels. The axes are generated using d3.extent and d3.scale.linear.
I've tried to de-conflict the point labels and the x-axis by extending the length of the y-axis, but the closest I've come to achieving this is by changing the original value of 0 to 30 in the following stanza:
var yext = d3.extent(data, d => d[1]);
var sy = d3.scale.linear()
.domain(yext)
.range([height, 30]) // flip y-axis
.nice();
The result is less than ideal, leaving an awkward gap instead of an intersection between the x and y axes:
What I want to achieve is something like this:
(Except I want to achieve this through code, rather than Photoshop).
Can anyone demonstrate a solution? (Plunker here)
Why don't you add a padding in the domain? Like:
.domain([yext[0] * 0.95, yext[1] * 1.05])
//less here---------^-- more here----^
Here is the plunker with that solution: http://plnkr.co/edit/rKArjn7DwQa9g1X5CaNW?p=preview
I'm working on a heatmap which basically plots the graph between taxIDs and KeywordNames from an external JSON. While I'm able to plot the values I see many blank spaces on the graph and clueless how I can plot them with the available data.
Here's the link to codesandbox: https://codesandbox.io/s/40mnzk9xv4
On the X-Axis I'm plotting the TaxIDs which are being calculated within the given range. I did try using the function rangeBands() but I get an error everytime.
Its the similar case with Y-Axis where I'am plotting the keywordIDs which are also being calculated within a range. I'm trying to print all the KeywordNames on Y axis and all taxIDs on the X-Axis and plot their corresponding spectracount on graph.
Please help me where have I gone wrong.
The output I'm looking for is something similar to this: https://bl.ocks.org/Bl3f/cdb5ad854b376765fa99
Thank you.
Some things to help you get you one your way:
First, your scales should use scaleBand(), not scaleLinear(), as they have discrete domains (i.e. categories of something, rather than continuous)
Second, your scale domains is taking every value of taxId and keywordName in your data as a possible value. But some values are repeated more than once. You need to be filtering them so you only have unique values. So your scale code should be:
const xValues = d3.set(data.map(d => d.taxId)).values();
const yValues = d3.set(data.map(d => d.keywordName)).values();
const xScale = d3.scaleBand()
.range([0, width])
.domain(xValues); //X-Axis
const yScale = d3.scaleBand()
.range([0, height])
.domain(yValues); //Y-Axis
Finally, your code that places the heatmap tiles needs to be calling the scale functions so it works out the position of each rect correctly:
chart.selectAll('g')
.data(data).enter().append('g')
.append('rect')
.attr('x', d => { return xScale(d.taxId) })
.attr('y', d => { return yScale(d.keywordName) })
That should get you most of the way there. You'll want to also reduce cellSize and remove the tickFormat calls on the axes as they are trying to convert your text labels to numbers.
I'm trying to create a simple scatter plot in d3 (similar to this one from matplotlib):
I use extent() to set the scale's input domain range.
xScale.domain(d3.extent(xvalues));
Using this approach results in some dots overlapping axises in d3 plot:
How to avoid axis overlapping and make a margin similar to the matplotlib's plot?
Input values vary, so simple increment / decrement of the extent() output doesn't look like a general solution.
In general, the best way of handling this is to call the scale's .nice() function, which will round the ends of the domain of the scale to nice values. In your particular case, this doesn't work, as the values are "nice" already.
In this case I would compute the extent of the domain and extend it by a fraction of that. For example:
var padding = (xScale.domain()[1] - xScale.domain()[0]) / 10;
xScale.domain([xScale.domain()[0] - padding, xScale.domain()[1] + padding]).nice();
In your matplotlib image, the dots are not overlapping and the x scale has negative value.
In d3:
var xScale = d3.scale.linear()
.domain([
d3.min(data, function(d) {
return d.val;
})-10, //so the domain is wider enough for the zero value
d3.max(data, function(d) {
return d.val;
}),
])
.range([height , 0])
I created a clock with d3 by making a bar chart updates as each second passes.
Typically I would set my yScale like so:
var yScale = d3.scale.linear()
.domain(maximumTime)
.range([svgHeight, 0]);
but this will leave an a blank bar at each new minute, hour, and day. To solve for this I've set the to go to -1 so something is always displayed, even when the value is zero.
var yScale = d3.scale.linear()
.domain(maximumTime)
.range([svgHeight, -1]);
Is there a more graceful way that I can display zero values?
Full example in a CodePen: http://codepen.io/agconti/pen/vEWZXb
I'm attempting to make a histogram using primarily time and date data, provided in a json file (along with other info) in this format: 2014-03-01 00:18:00. I've looked at http://bl.ocks.org/mbostock/3048450 as an example, but I haven't managed to crack it. The key part seems to be this:
var data = d3.layout.histogram()
.bins(x.ticks(20))
(dataset.timestamp);
When I view my code in the browser it gives "TypeError: data is undefined", and refers to d3.v3.js line 5878.
Assuming I fix that error, the next place it may stumble is the axis formatting:
var formatDate = d3.time.format("%y-%m-%d %h:%m:%s");
var xAxis = d3.svg.axis()
.scale(x)
.orient("bottom")
.tickFormat(formatDate);
Will the format syntax correspond correctly to my time and date format?
I don't think that the histogram layout accepts non-numeric input (I could be wrong). One option would be to convert the dates into the number of milliseconds since 1970 by parsing and using getTime():
var formatDate = d3.time.format("%Y-%m-%d %H:%M:%S");
var data = d3.layout.histogram()
.bins(x.ticks(20))
.value(function(d) {return formatDate.parse(d.timestamp).getTime()})
(dataset);
You'll need to make sure that the x scale (as in x.ticks(20)) has a domain based on the millisecond data.
So I figured it out, thanks to the helpful answers here. First off, the json data is acquired:
d3.json("inkmarch.json", function(error, json) {
if (error) return console.warn(error);
dataset = json;
Then I specify a formatting variable specific to my date formatting:
var formatDate = d3.time.format("%Y-%m-%d %H:%M:%S");
I map the re-formatted dates to another var:
mappedSet = (dataset.map(function(d) {return formatDate.parse(d.timestamp).getTime()}))
And finally I can make a proper histogram dataset.
var data = d3.layout.histogram()
.bins(x.ticks(31))
.value(function(d) {return formatDate.parse(d.timestamp).getTime()})
(dataset);
(31 ticks since this is monthly data from March). I did away with the d3.time.scale since they were integers anyway, so my scales look like such:
var x = d3.scale.linear()
.domain([d3.min(mappedSet), d3.max(mappedSet)])
.range([0, width]);
var y = d3.scale.linear()
.domain([0, d3.max(data, function(d) { return d.y; })])
.range([height, 0]);
(Still not sure what the function in the y domain does). The svg and bar vars look exactly like the example I linked in my question post, as do the 'rect' and 'text' bars (except I changed width of rect and x position of the text to fixed values since the example formulas were giving hideous negative values).
Thanks again for the help, I'm sure I'll have more d3 questions in the near future.