Scilab Website | Contribute with GitLab | Mailing list archives | ATOMS toolboxes
Scilab Online Help
2024.0.0 - Français


stackedplot

plot multiple timeseries on time axis

Syntax

stackedplot // demo
stackedplot(ts)
stackedplot(ts1, ..., tsN)
stackedplot(..., vars)
stackedplot(..., LineSpec)
stackedplot(..., LineSpec, GlobalProperty)
stackedplot(..., "CombineMatchingNames", %f)
stackedplot(..., Name, Value)
f = stackedplot(..)

Arguments

ts, ts1, ..., tsN

timeseries object

vars
  • strings vector containing to variables names of timeseries

  • column index vector corresponding to position in variables names timeseries

  • cell of strings: each element can contain one or multiple variables names.

LineSpec

This optional argument must be a string that will be used as a shortcut to specify a way of drawing a line. We can have one LineSpec per y or {x,y} previously entered. LineSpec options deals with LineStyle, Marker and Color specifiers (see LineSpec). Those specifiers determine the line style, mark style and color of the plotted lines.

GlobalProperty

This optional argument represents a sequence of couple statements {PropertyName,PropertyValue} that defines global objects properties applied to all the curves created by this plot. For a complete view of the available properties (see GlobalProperty).

Name, Value

'Title': string that contains the title of plot.

'DisplayLabels': matrix of strings that contains the names of Y-label for each axe. Default value: variables names of timeseries

'LegendLabels': matrix of strings that contains the names of timeseries.

f

handle of the created figure.

Description

The stackedplot function plots the variables of timeseries on the same time basis (x-axis). All timeseries must have the same time basis. If two or more timeseries have common variables, their curves will be in the same graph. By specifying CombineMatchingNames to %f, there will be one graph per variable for each timeseries.

stackeplot(..., vars) plots only variables of timeseries specified by vars. vars can be

  • a vector of strings containing variables names to plot.

  • a vector of indexes ccorresponding to position of variables names in timeseries.

  • a cell containing variable names in matrices to plot them in same graphs. ie: {"var1", ["var2", "var3"]} will display var1 in a graph and var2 and var3 in another one.

stackedplot(..., LineSpec, GlobalProperty) customizes the line style, marker or color of polylines. This arguments are the same as the plot function.

stackedplot(..., "Title", str) adds a plot title.

stackedplot(..., "DisplayLabels", str) specifies the labels for the y-axis.

stackedplot(..., "LegendLabels", str) adds a legend for each plot. str is a vector of size N timeseries input. The legend can be used to specify the name of timeseries.

Enter the command stackedplot to see a demo.

Examples

stackedplot(ts)

n = 100;
t = datetime(2000, 1, 1) + caldays(1:n)';
y = floor(10 * rand(n, 3)) + 10;
ts = timeseries(t, y(:, 1), y(:, 2), y(:, 3), "VariableNames", ["Time", "Var 1", "Var 2", "Var3"]);
stackedplot(ts);

stackedplot(ts1, ts2)

n = 100;
t = seconds(1:n)';
y1 = floor(10 * rand(n, 3)) + 10;
ts1 = timeseries(t, y1(:, 1), y1(:, 2), y1(:, 3), "VariableNames", ["Time", "Var 1", "Var 2", "Var3"]);

y2 = floor(10 * rand(n, 2)) + 20;
ts2 = timeseries(t, y2(:, 1), y2(:, 2), "VariableNames", ["Time", "Var 1", "Var 2"]);
stackedplot(ts1, ts2);

stackedplot(ts1, ts2, "CombineMatchingNames", %f)

t = datetime(2022, 1, 1):datetime(2022, 8, 31);
n = size(t, "*");
y1 = floor(10 * rand(n, 2)) + 10;
ts1 = timeseries(t, y1(:, 1), y1(:, 2), "VariableNames", ["Time", "Result_1", "Result_2"]);

y2 = floor(10 * rand(n, 2)) + 20;
ts2 = timeseries(t, y2(:, 1), y2(:, 2), "VariableNames", ["Time", "Result_1", "Result_2"]);
stackedplot(ts1, ts2);
stackedplot(ts1, ts2, "CombineMatchingNames", %f);

stackedplot(ts1, ts2, vars)

t = datetime(2022, 1, 1):datetime(2022, 8, 31);
n = size(t, "*");
y1 = floor(10 * rand(n, 3)) + 10;
ts1 = timeseries(t, y1(:, 1), y1(:, 2), y1(:, 3), "VariableNames", ["Time", "Result_1", "Result_2", "Result_3"]);

y2 = floor(10 * rand(n, 3)) + 20;
ts2 = timeseries(t, y2(:, 1), y2(:, 2), y2(:, 3), "VariableNames", ["Time", "Result_1", "Result_2", "Result_3"]);
stackedplot(ts1, ts2, ["Result_1", "Result_3"]);

stackedplot(ts1, ts2, [1 3]);

stackedplot(ts1, ts2, {"Result_1", ["Result_2", "Result_3"]});

stackedplot(ts1, ts2, "LegendLabels", str, "DisplayLabels", str2)

t = datetime(2022, 1, 1):datetime(2022, 8, 31);
n = size(t, "*");
y1 = floor(10 * rand(n, 3)) + 10;
ts1 = timeseries(t, y1(:, 1), y1(:, 2), y1(:,3), "VariableNames", ["Time", "Result_1", "Result_2", "Result_3"]);

y2 = floor(10 * rand(n, 3)) + 20;
ts2 = timeseries(t, y2(:, 1), y2(:, 2), y2(:, 3), "VariableNames", ["Time", "Result_1", "Result_2", "Result_3"]);
stackedplot(ts1, ts2, {"Result_1", ["Result_2", "Result_3"]}, "LegendLabels", ["t1", "t2"], "DisplayLabels", ["res1", "res2 - res3"]);

See also

  • timeseries — create a timeseries - table with time as index
Report an issue
<< Sgrayplot 2d_plot 3d_plot >>

Copyright (c) 2022-2023 (Dassault Systèmes)
Copyright (c) 2017-2022 (ESI Group)
Copyright (c) 2011-2017 (Scilab Enterprises)
Copyright (c) 1989-2012 (INRIA)
Copyright (c) 1989-2007 (ENPC)
with contributors
Last updated:
Tue Oct 24 14:34:14 CEST 2023