Parallel coordinates with d3.js - javascript

I need to to represent a Parallel Coordinates with d3.js. First of all I have no idea if what I think is possible to achieve (and if the Parallel Coordinates is the right chart).
I will explain my idea: I take data from a database and I expose them in a JSON and I store them in an array of objects (with JavaScript).
This is an example of the data http://pastebin.com/DZcMqDMc.
I would like to represent along the abscissa axis years (though there are years repeating themself, as you can see from data example), while along the ordinate axis values of those years (values are in percent, ranging from 1 to 100).
I would like to represent two lines according to "value1" and "value2" property in the JSON file.
Is it possible? Is Parallel Coordinates the right chart?
The main problem I have right now is that I do not understand how to set right the two domains (abscissa and ordinate).
I am basing on the example Parellel Coordinates of Bostock.
For abscissa I am thinking something like that:
x.domain(
d3.extent(test,
function(d) {
return d.years;
}
)
);
It makes sense or?

Try a multi line chart.That might suit your need.
I am not sure what you are trying.

Simply do a line chart, and produce the vertical lines by formatting the ticks to go from 0 to height ( in your var xAxis code include .tickSize(0-height)). You will have to pick the right number of ticks, as in ticks(), so you just get the lines where you want them.

Check out Parcoords, a d3-based parallel coordinates library. For compatibility with d3 v5 see https://github.com/BigFatDog/parcoords-es which is based on the original Parcoords library (https://github.com/syntagmatic/parallel-coordinates) which relies on an outdated version of d3.
For examples and sample code, check out the following link: http://syntagmatic.github.io/parallel-coordinates/.

Related

Is there any d3 function like scalePoint but for two dimensions?

I want to place clusters like in this block:
https://bl.ocks.org/SpaceActuary/d6b5ca8e5fb17842d652d0de21e88a05
However, I want the clusters to be on a 2d grid rather than in a line and specify range in each dimension so that I can have e.g. a short but wide lattice.
I tried playing with scalePoint range but couldn't get it to return an array.
I could manually compute positions, but would be a lot cleaner if there's already a built-in way to do this in d3.

NVD3 - configuring ticks on axis

I have a nvd3 line chart which displays a time series and can't get the ticks on the x axis right.
For longer time spans, it works as expected. But for shorter time spans (here: 12/31/05 to 01/01/06), the same date is displayed for multiple ticks:
Please have a look at the code for this chart on JSFiddle
I want the chart to only display ticks at data points, and not in between. Is that possible with a line chart? From my understanding, it is possible with d3, but I can't figure out if this functionality is exposed by nvd3.
I've tried explicitly setting the number of ticks with chart.xAxis.ticks() without success. The only thing that has any effect is explicitly setting the tick values with chart.xAxis.tickValues([...]), but I would prefer not having to calculate them myself.
The way to solve this in general is with custom multi-scale time formats. Note that this example itself will not work with NVD3 because it uses an older version of D3, the examples below will though.
The problem in your case is that the ticks aren't "clean" divisions of time and if you apply a multi-scale format, you get something like this. It always shows the more fine-grained format because anything else would involve a loss of precision.
You can however use a simple heuristic to show the date instead of the time if the hour is less than 3, which works reasonably well in your case. See here for an example. The proper way to do this would be to make your ticks clean divisions.
Which brings us to your actual question. There's no other way than to explicitly set .tickValues() for what you want to do, but you can compute the x positions in your data quite easily:
var xvalues = [],
tmp = data.map(function(e) {
return e.values.map(function(d) { return d[0]; });
});
xvalues.concat.apply(xvalues, tmp);
The code is not the prettiest because it's a nested structure, but fairly straightforward. Using this, you can set your tick values explicitly, full example here.

Plotting extra linear data on D3 js graph

Check out this fiddle, using D3 and plotting a line graph against dates.
http://jsfiddle.net/T546B/172/
I want to plot some further data on my graph and want to know if its possible, I would ideally like the graph to look like below:-
The extra data is linear and doesn't have a price value and I want it to be included somewhere near the middle of the graph. The data array would be in a format along the lines of:-
var eventArray = [[startdate, enddate, name]];
I was basically wondering if this is possible, plotting two types of data, using different SVG elements on one graph. - not sure how to approach this problem. Any help appreciated!
you just need to set up a second line generator with the coordinates that you want for x and y

d3.js scatter matrix with brushing - scale error

I try to reuse the example of 'scatter matrix with brushing': http://bl.ocks.org/mbostock/4063663
It seems that the code is not directly reusable with another csv. Scales seems to be somehow hard coded or so: i change the csv by adding 10 to 75% of the first column values, and the xscale is not directly updated.
To visualize the problem, see the fork of the mbostock gist: http://bl.ocks.org/fdeheeger/7249196
I cannot figure out where/how the scale is computed or updated in the javascript code.
Any advice from a d3 expert?
The scales are computed dynamically -- the problem is that the numbers in the CSV are parsed and processed as strings and not numbers. This is also the case in the original block, but there it doesn't matter because the ordering of the strings is the same as the ordering of the numbers.
All you need to do to fix this is parse the strings to numbers:
domainByTrait[trait] = d3.extent(data, function(d) { return +d[trait]; });
The plus makes all the difference here. Complete example here.

Charting thousands of points with dojo

I need to plot thousands of points, perhaps close to 50,000 with the dojo charting library. It works, but it's definitely very slow and lags the browser. Is there any way I can get better performance?
EDIT:
I solved by applying a render filter to the data. Essentially, I have a new item parameter called "render" which is set to false by my json source if the point is expected to overlap others. My DataSeries then queries for all points where render:true. This way all of the data is there still for non-visual sources that want all of the points, while my charts now run smoothly.
Psuedocode:
def is_overlapped(x, y, x_round, y_round)
rounded_x = round(x, x_round)
rounded_y = round(y, y_round)
hash = hash_xy(rounded_x, rounded_y)
if(#overlap_filter[hash].nil?)
#overlap_filter[hash] = true
return false
end
return true
end
x_round and y_round can be determined by the x and y ranges, say for example range / 100
I know this isn't probably exactly the answer you're looking for, but have you considered simply reducing the number of points you are plotting? I don't know the specific function of the graph(s), but I'd imagine most graphs with that many points are unnecessary; and no observer is going to be able to take that level of detail in.
Your solution could lie with graphing techniques rather than JavaScript. E.g. you could most likely vastly reduce the number of points and use a line graph instead of a scatter plot while still communicating similar levels of information to your intended target.

Categories