ScottPlot.NET
GitHub Repo stars

How to plot data on a DateTime axis

C# can represent time with the DateTime type, but all of ScottPlot’s methods prefer double arrays. This page reviews how to plot data with a time axis using ScottPlot.

💡 TLDR: Convert DateTime to double using DateTime.ToOADate() before plotting, then call XAxis.DateTimeFormat(true) to display axis tick labels in the proper format.

Scatter Plot

  • If time points are not evenly spaced, the Scatter Plot should be used (otherwise a Signal plot should be considered).
  • Use LINQ and DateTime.ToOADate() to convert DateTime[] to double[] before plotting
  • Call XAxis.DateTimeFormat(true) to indicate axis tick labels should be formatted as DateTime values (instead of the default numeric format)
// generate sample data (arrays of individual DateTimes and values)
int pointCount = 100;
Random rand = new Random(0);
double[] values = ScottPlot.DataGen.RandomWalk(rand, pointCount);
DateTime[] dates = Enumerable.Range(0, pointCount)
                              .Select(x => new DateTime(2016, 06, 27).AddDays(x))
                              .ToArray();

// use LINQ and DateTime.ToOADate() to convert DateTime[] to double[]
double[] xs = dates.Select(x => x.ToOADate()).ToArray();

// plot the double arrays using a traditional scatter plot
var plt = new ScottPlot.Plot(600, 400);
plt.AddScatter(xs, values);

// indicate the horizontal axis tick labels should display DateTime units
plt.XAxis.DateTimeFormat(true);

// add padding to the right to prevent long dates from flowing off the figure
plt.YAxis2.SetSizeLimit(min: 40);

// save the output
plt.Title("Scatter Plot with DateTime Axis");
plt.SaveFig("../../../datetime-scatter.png");

Source code for this application is available on GitHub

Signal Plot

  • Signal plots offer a huge performance increase over Scatter Plots
  • Signal plots calculate X position from sample rate and offset
  • The value of sample rate is the number of samples per day:
    • daily: sampleRate = 1
    • hourly: sampleRate = 24
    • minute: sampleRate = 24 * 60
    • second: sampleRate = 24 * 60 * 60
    • millisecond: sampleRate = 24 * 60 * 60 * 1000
// generate sample data (array of values with a single starting DateTime)
int pointCount = 100;
Random rand = new Random(1);
double[] values = ScottPlot.DataGen.RandomWalk(rand, pointCount);
DateTime firstDay = new DateTime(2016, 06, 27);
double samplesPerDay = (24 * 60); // one sample per minute

// plot the double arrays using a traditional scatter plot
var plt = new ScottPlot.Plot(600, 400);
var sig = plt.AddSignal(values, sampleRate: samplesPerDay);
sig.OffsetX = firstDay.ToOADate();

// indicate the horizontal axis tick labels should display DateTime units
plt.XAxis.DateTimeFormat(true);

// add padding to the right to prevent long dates from flowing off the figure
plt.YAxis2.SetSizeLimit(min: 40);

// save the output
plt.Title("Signal Plot with DateTime Axis");
plt.SaveFig("../../../datetime-signal.png");

Notes

  • 👉 The ScottPlot Cookbook has additional information and code examples demonstrating advanced DateTime tick label functionality

  • Axes have additional options for tick formatting, including the ability to define format strings for the date-to-string conversion, as well as the ability to indicate which culture to use (e.g., different cultures periods, hyphens, or dashes to separate month and year).