Prepare a figure

[dmkahler]

How do I prepare a figure for publication?

[dmkahler]

Most publications require figures to be prepared in vector format. In general, there are two types of images, vector and raster. A raster image is like a photograph; it is essentially a table of locations with the color in each cell. You have seen the limitation of this type of image when you zoom in too far and the image becomes pixelated. Vector formats do not have this problem. A vector image is a set of instructions for the computer to redraw a perfectly crisp version of your image at any zoom level. The most common file type is encapsulated postscript (.eps).

[dmkahler]

Luckily, most scientific software packages will export to .eps. One standard package is R, which can easily be used with RStudio. See my lab wiki for more on R and RStudio installation.

I highly recommend the ggplot2 library in R, which stands for the grammar of graphics. You can use this package to produce really nice figures. There are a lot of controls, so take your time learning some of the different options; a good place to start is the tidyverse documentation.

If you have data already in R (to use ggplot, you must have your data in a data frame), try these commands:

library(ggplot2)
library(latex2exp
fig <- ggplot(dataframe) +  
      geom_point(aes(x = x_var, y = y_var)) + 
      labs(x = "Date", y = TeX('Power ($W/m^2$)')) + 
      xlim(0,10) + 
      ylim(0,10) + 
      theme(panel.background = element_rect(fill = "white", colour = "black")) + 
      theme(aspect.ratio = 1) + 
      theme(axis.text = element_text(face = "plain", size = 12))
ggsave(filename.eps, plot = fig, device = "eps")

For fancier labels, check out the library, latex2exp. This includes the tex command shown for the y-axis label.

[profhuster]

I use python and the matplotlib library from a jupyter notebook. Like probably most ways, there is a learning curve. Here is a code example and the figure. You can change the EPS format in the last line by using Sound_Speed_vs_Temperature.svg
for SVG format (a vector), Sound_Speed_vs_Temperature.png (Raster image format), etc.

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
temperature = np.linspace(0, 40, 401)
sound_speed = 331.3 * np.sqrt(1 + temperature / 273.14)
fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(temperature, sound_speed)
ax.set_title("Sound Speed versus Temperature")
ax.set_xlabel("Temperature (°C)")
ax.set_xlim(0,40)
ax.set_ylabel("Speed of Sound (m/s)")
ax.set_ylim(330,355)
ax.grid()
fig.savefig("Sound_Speed_vs_Temperature.eps")

[corcoted]

For people looking for a pre-packged solution, Veusz (https://veusz.github.io/) is an open source app (built with python) that works similar to commercial software like Origin or Kaleidograph. It has all of the bells and whistles to make production-quality figures.