Plots Publication Format Figures Multiple. 2D Plots. K. Cooper 1. 1 Department of Mathematics. Washington State University.

Transcription

1 2D Plots K. 1 1 Department of Mathematics 2015

2 Matplotlib The most used plotting API in Python is Matplotlib. Mimics Matlab s plotting capabilities Not identical plot() takes a variable number of arguments... Arrays of abscissae and ordinates Optional strings describing line and marker format Optional keyword=value arguments to describe changes from defaults. Easier to see in examples than to describe

3 Plot() x=arange(-1.,1.01,.01) plot(x,sin(pi*x)) show()

4 Plot() x=arange(-1.,1.01,.01) plot(x,sin(pi*x),x,cos(pi*x)) show()

5 Plot() x=arange(-1.,1.01,.01) plot(x,sin(pi*x), r:,x,cos(pi*x)) show()

6 Matlab Linespecs Line types: - solid; -- dashed; : dotted; -. dash-dot Colors: r red; g green; b blue; c cyan; m magenta; y yellow; k black; w white Markers: o circle; + plus symbol; * asterisk;. point; x ex; s square; d diamond; ˆ upward triangle; v downward triangle; < leftward triangle; > rightward triangle; p pentagram; h hexagram

7 Keywords x=arange(-1.,1.01,.01) plot(x,sin(pi*x),linewidth=2) plot(x,cos(pi*x),color= orange ) show()

8 Titles, Labels... plot(x,sin(pi*x),linewidth=2) plot(x,cos(pi*x),color= #f0a020 ) xlim(-1.,1.) ylim(-1.2,1.2) xlabel( x ) ylabel( f(x) ) title( Trig Fcts ) show()

9 Enhancements Graphics intended for publication require some extra thought. They will probably be reduced in size... Fonts must be larger Fewer tick marks, less busy There will be a caption we do not need/want a title On the other hand, the figure must be somewhat self-explanatory: clear labels, important features marked.

10 Tick Marks and Labels. xlabel( x ) ylabel( f(x) ) ax = gca() ax.set_xticks([-1.,0.,1.]) ax.set_xticklabels([ -1, 0, 1 ],fontsize=18) ax.set_yticks([-1,0,1]) ax.set_yticklabels([ -1, 0, 1 ],fontsize=18) show()

11 Legend and LaTeX rcparams.update( font.size :14, text.usetex : True) fct_label = r $\sin(\pi x)$ drv_label = r $\frac{d}{dx}\sin(\pi x)/\pi$ plot(x,sin(pi*x),linewidth=2.0,label=fct_label) plot(x,cos(pi*x),color= #f0a020,label=drv_label) legend(loc=4)

12 Legend Location We can move the legend more or less where we want relative to the figure bounding box. legend(bbox_to_anchor=(.8,.25)) The numbers are x and y coordinates relative to the figure not the axes, So we could even put it outside the axes.

13 Annotations Add a vertical line to indicate the relationship between the points on the graph of the function and those on the graph of the derivative. plot([0,0],[0,1], -,color= # ) plot([0],[0], bo,[0],[1], bo ) Matplotlib is happy to make vertical lines.

14 Annotations Now insert text to explain the points. string to appear on plot annotate(r $\text{slope}=1$, xy=(0,0),\ xytext=(-.45,-.05),\ coords of point of interest coords arrowprops={ facecolor : black, width :0.5,\ of text shrink :0.1, headwidth :5}) annotate(r $\frac{d\,\sin}{dx}(0\pi)/\pi=1$,xy=(0,1 xytext=(-.6,.95),\ arrowprops={ facecolor : black, width :0.5,\ shrink :0.1, headwidth :5}) dictionary of arrow properties

15 Annotations

16 Annotations

17 Image Formats There are several image formats, with somewhat different properties. JPEG Compressed using Fourier series (JPEG1) or wavelets (JPEG2). Lossy: if you use this for diagrams, you may see errant pixels near what should be clean lines. PNG Compressed using a simple description. Lossless: your lines will be clean, but if you change the size, information will be lost. GIF An old lossless compressed format. Do not use this. Bitmap A lossless uncompressed format. Do not use this.

18 Image Formats For your own use, save images any way you want. For presentations, it is probably fine to use PNG, and allow resizing. For publication, it is probably best to use PNG in a format that does not require resizing. I.e. it should appear in your paper without e.g. a width= modifier.

19 Figures There are three basic parts of a plot. 1. Data series the actual curves or points we plot 2. Axes the plotting area, including tick marks and labels 3. Figures this is the entire area that might hold several axes. Each part has its own properties.

20 Figures We can make figures any time we want using the figure() command. It is best to assign a figure handle to a variable: e.g. f1 = figure(). We can get the handle for the current figure any time using e.g. f_current = gcf() Only needs one show() to show all figures.

21 Subplot The subplot command allows us to make multiple plots in a given figure. The plots then save as one image. subplot(1,2,1) plot(x,sin(pi*x)) xlim(0.,1.) ylim(-.2,1.2) subplot(1,2,2) plot(x,cos(pi*x), g ) xlim(0.,1.) ylim(-1.1,1.1) show()

22 Subplot We can change the aspect of the figure just by changing the figure window drag with the mouse. subplot(1,2,1) plot(x,sin(pi*x)) xlim(0.,1.) ylim(-.2,1.2) subplot(1,2,2) plot(x,cos(pi*x), g ) xlim(0.,1.) ylim(-1.1,1.1) show()

23 Extra Axis We can make additional axes on which to plot details of our existing plot... Coords in add_axes are (xleft,ylower,width,height) relative to the figure.. myfig.add_axes([.65,.4,.2,.2]) plot(x,sin(x)) plot(x,x*(1.+x*x*(-1./6.+x*x*(1./120.-x*x/(120*6.*7 xlim(2.3,2.5) ylim(.6,.7) ax.arrow(1.9,.25,.5,.32,\ facecolor= black,\ length_includes_head=true,\ head_width=.04)

24 Other 2D Plot Types semilogx(), semilogy(), loglog() hist() histogram streamplot() stream lines bar() bar charts pie() pie charts polar() polar plots contour() contour plots