This question is kind of conceptual, so if this isn't the right place for this, I apologize.
I have an existing scatterplot that has continuous Y-axis data and used to have continuous X-axis data. My X-axis data has now changed to semi-continuous, it's a set series of number (cost values), like [200, 450, 600, 900, 1400]. There are about 250 'points' on the scatter-plot, but now it seems strange because the X values are not continuous, and they create vertical columns of points.
Without creating an entirely new graph, is there something I can add or do to these data points to better represent them? I was thinking something along the lines of a box-whisker, but a box-whisker wouldn't work for these values.
Edit Question Clarification:
Is there a type of chart or adjustment that I can make to this chart to better display these values as unique, interesting points of data?
Related
I'm using chart.js 2.9.4 and the ng2-charts wrapper for Angular. I'm trying to display incoming live data, but have trouble configuring the chart so that the ticks/axis don't stretch beyond the data in the chart. In other words, I want the chart data points to fill the entire width of the chart grid.
StackBlitz showing my issue.
If you keep looking at the chart as data is added, you see that most of the time the ticks stretch beyond the last point in the chart:
The only solution I could come up with is overwriting the max value of the ticks on the x-axis each time new data is added to the chart: options.scales.xAxes[0].ticks.max = x;. Uncomment line 68 in the StackBlitz in order to apply this. This solves my problem but introduces another. Sometimes as data gets added, the last two ticks overlap:
I've tried experimenting with various parameters of the scales and ticks options (bounds,distribution,stepSize,source,autoSkip,autoSkipPadding) with no results. Is there any combination of configuration parameters to solve my issue ?
You may want to add delay at streaming plugin, the config will look something like:
scales: {
x: {
type: 'realtime', // x axis will auto-scroll from right to left
realtime: { // per-axis options
delay: 1000, // delay of 1000 ms, so upcoming values are known before plotting a line
pause: false, // chart is not paused
ttl: undefined, // data will be automatically deleted as it disappears off the chart
}
},
I have tested this on your code, and it seems to work as expected, you may want to tune some parameters as needed.
I am having problems controlling the Y-Axis range of a highcharts graph. It seems like highcharts likes nice round numbers. When my data passes or is close to certain thresholds, the Y-Axis range can expand a lot which effectively compresses all the plot points downward.
Here is a jsfiddle that illustrates the problem I am having:
https://jsfiddle.net/shannonwrege/z8h5eork
The relevant code for this post is this:
chart.yAxis[0].setExtremes(0, max, true, false);
Keep in mind that I don't know what the data will look like in advance, so I must dynamically modify the Y-Axis range. Right now I am using the setExtremes because of other suggestions I've read on stackoverflow.
The maximum y-value of the data in the first two charts is 99. You'll notice that the y-axis is set at 150 in the first chart where the range is automatically calculated and 100 in the second chart where I specifically set the extreme values. The look of the 2nd chart is what I want. So it seems like setExtremes(0,99,true,false) should do the trick, but it actually doesn't.
In the 3rd chart I changed the data so that the maximum y-value of the data is 101, and I called setExtremes(0,101,true,false). You'll note that the y-axis is now back to 150.
Ideally I want the scale of the graph to be capped on the maximum value to limit the about of extra white space. I want to see all of the data, but I don't necessarily care about the y-axis displaying a maximum band that is greater than the max data value. In this case, I would be happy with the y-axis displaying 100 on the axis and some points over but still visible.
Does anyone know how to make this work?
I ended up using the endOnTick parameter to solve this problem. Adding the following line to the yAxis configuration parameters did exactly what I wanted:
endOnTick: false,
Here's the updated Fiddle showing the results.
https://jsfiddle.net/shannonwrege/z8h5eork/3/
All of the charts look pretty good in my opinion (even the one where the yAxis range was auto calculated).
You will need to read the data and then round up to set the idealMax
var chart,
idealMax = 0; // init the max value
// Read the data to find the highest value
for (i=0;i < (options.series[0].data).length; i++ ){
if (options.series[0].data[i][1] > idealMax) {
idealMax = options.series[0].data[i][1];
}
}
// Round the max to the nearest 10
idealMax = Math.round(idealMax / 10) * 10;
options.yAxis.tickPixelInterval = idealMax/10;
Highcharts.chart('container1', options);
chart = $('#container1').highcharts();
chart.yAxis[0].setExtremes(0, idealMax, true, false);
Updated Fiddle
Is there any workaround to have tickMarkPlacement set as "between" on a datetime Axis? I am aware is not supported by the API but I was hoping to find some sort of hack/plugin.
You can fudge it with the x axis label's x property.
Example:
http://jsfiddle.net/jlbriggs/3qtZr/36/
[[update after comments:
if you mean that you want the data points to also line up 'between', then there isn't a good easy way.
My approach would be
1) make sure there is only one data point per axis tick, ideally
2) adjust the x value of the data point to push it to the right in a way that corresponds with the label offset
3) adjust tooltip formatters to correct the date for display
or, 4) just go with categories
However, if you can explain why you want to do this, what effect you're going for, perhaps there's more that can be done.
{{further updates:
After playing around a little more, I found another way to fudge the data point placement, though I am unclear whether you need that.
Example here:
http://jsfiddle.net/jlbriggs/3qtZr/39/
It uses the pointPlacement property.
The catch is that the pointPlacement property doesn't work if there is not a columns series present with as many data points as the line series, it seems.
So this method adds and hidden dummy column series in order to make the pointPlacement property affect the line series.
Not elegant, but it beats having to adjust the data values and then re-adjust them in the formatter.
I want to create a histogram in Highcharts. The bin series has about 8 elements. The series for the the distribution curve has about 200 elements. Since Highcharts infers the xAxis from the number of elements in the series, the xAxis stretches out to 200. How do I get the curve series to fit to the bin series on the xAxis?
I would suggest using 2 x axes for this. It is far easier than trying to make the points match on a single axis.
See my example here:
http://jsfiddle.net/FnhRV/19/
Well, first I advice to get familiar with Highcharts docs/tutorials. Like this one.
In general, you can manage distance between points, it's called pointInterval, for example: http://jsfiddle.net/Dd9Py/1/
When you have 8 columns, on xAxis you should have scale - according to pair [x,y] of values.
Another solution is to use two different xAxis, one for column and one for spline. Example: http://jsfiddle.net/Dd9Py/2/
Looking at this jsFiddle. Things I notice is that your xAxis is linear - meaning that each point is plotted consecutively. So, since your bar series only has 8 points they are plotted int he first 8 positions. Your curve contains 200 points and are also plotted first come first served. You need to link up your xAxis so that each series is linked. What are your xAxis increments/categories?
I have a chart here: http://jsfiddle.net/wergeld/bx82z/
What I need is for the stacked bars and the scatter points to line up on each other based on the data X-element (county name in this case).
Getting the bars to stack was simple. But as you can see in the example the scatter points show up in the middle of the chart - not on the same x-value as the others.
I suppose I could create a category list (using the county names) but there may be cases where I do not have all 3 data elements for all counties - how to handle that scenario?
I did great the category axis and it is still doing the same thing: http://jsfiddle.net/wergeld/YckwW/
You were not defining the series correctly to achieve what you were wanting. See updated jsfiddle for what I think you are looking for.
In the cases where you are missing a piece of data for one of the countries you could always use null, which simply means there is nothing to display for that data point. See Missing Data Sample for an example. It is case sensitive so make sure that null is all lower case.