Virtually all scientific reports and articles deal with numbers of some sort, and what better way to represent large amounts of data than on a well made graph? Unfortunately not all graphs are created equal; this post will introduce the basics of graph building and common pitfalls to avoid.
This post is more for the students who are transitioning from high school to college, and will address the basics, from formatting to plotting – so they don’t have to learn it the hard way! The programs that I use are ProFit and SigmaPlot, but the tips here will generally apply to any software out there.
The type of graph depends on the data that you have, but generally at this level you will get a range of values for both dependent and independent variables – so scatter plots are the name of the game. Let’s load up a sample data set:
Your graph should always have context! Whether the title of your graph is placed on top or in the figure description it has to make sense to the reader. A general rule of thumb is to name the graph:
‘[y-axis] vs [x-axis] of/for [system]’
This will give the reader ample description without cluttering the title with information. It is important to name it [y-axis] vs [x-axis] and not the other way round! This is as you are portraying y (dependent variable) as a function of x (independent variable).
UNITS ON AXES. Another thing you might want to include is a figure legend if you have multiple plots. A neat way to include it is within the empty space on the graph – this will make everything look neater, but outside of it works just as well. Make sure it’s easy to match the plot to the legend (using colours, shapes, etc).
Once you have the texty bits under control, you have the option to include gridlines. Gridlines are obviously more important if you are trying to highlight key values within your plots. Be careful not to go overboard; most graphing programs will let you adjust the intervals between the lines. And please please please, if you are going to use gridlines use them for BOTH x and y axes.
Another thing you may have noticed is the axis ticks, that I prefer to use facing into the graph. You may want to use outward facing ticks if you have plot points very close to the axes.
Best Fit Lines
You might be under the impression that a graph isn’t complete without a line of best fit running through as many points in the graph as possible, but this isn’t the case. This is especially true for systems that don’t follow a linear or log/exp pattern, where it is better to leave the plots on their own than try to fit them under a 7th order polynomial expression and extrapolate the data incorrectly.
There you have it, a quick guide to good graphing practices. Here’s one I used in my thesis where instead of a header title, I included a description in the figure caption.
Happy graphing! If you have any suggestions for more tutorial posts, please contact us!