Some graphs (I’m not sure what to call them) have a time line running horizontally and another line (quantity) running vertically. What do you call such graphs and what is the rule that states that time lines run horizontally (if that’s even true) and lines measuring quantity run vertically (if that’s even true)? I haven’t been able to find out much about these kinds of graphs. I know there is a logical reason for the structure of the graph, but I don’t know what it is. I look forward to your feedback.
davidmich
Is there some ‘rule’? I never worried about it if there is. I guess that what you describe might be a bar graph.
Most places I worked at had some kind of house style, or at least a common way of presenting data. Lacking that, I would choose whichever type of graph suited what I was trying to present. In general it is the vertical axis that has the greatest impact. Scale is important - You can easily smooth out, or exaggerate differences by choosing the right scale.
Google ‘graphs’ and you can find loads of resources, from the very basic to the highly sophisticated.
Conventionally, you have the independent variable on the horizontal axis and the dependent variable on the vertical. If one of the variables is time, it is almost always taken as the independent.
I’d call it a “convention” rather than a “rule”. It’s like a map wherein the convention is that north is up.
In a two-dimensional graph, the convention is that the independent variable (in this case, time) is on the horizontal axis and the dependent variable (e.g. widgets-per-week) is on the vertical scale. There’s no “rule” that increasing values go from left-to-right or from bottom-up. Since that’s the way most graphs are drawn, it makes it easier for the viewer.
This.
It’s easy enough to imagine time as the dependent variable though, eg “road distance vs driving time”.
Additionally, the type of graph usually helps indicate some detail of the variable(s).
In the example of distance vs. time, time is a changeable quantity. In something that tracks, say quantity vs. calendar time (or hourly) that time passes regardless of the value you measue - time is normally independent.
In variables where the current value is tied to the previous value (exchange rate, price of gold, etc.) a line graph is usually appropriate, you are following a trend. In a business where each instance is relatively independent (i.e. orders this month, number automobile accidents, etc.) a bar graph makes more sense. Each quantity stands on its own.
Bar graphs that extend from the y-axis horizontally are not unusual, but timelines, it seems, tend to go horizontal on the theory “we could extend this graph further”. It’s just convention, but the space to the right tends to be blank, while the space below fills with more text right away.
One of the best sources to answer this sort of question is Edward Tufte’s “The Visual Display of Quantitative Information.” It’s one of those rare books about a dry, scientific field that’s straight up fascinating and fun to read.
A common exception that I’ve seen: If the independent variable is altitude, it is often put on the vertical axis, just so “up means up”.
You’ll also sometimes see the independent variable as the vertical axis just because it fits on the page better, as with XKCD’s graph of the makeup of Congress. In that particular case, it also allows left to mean left and right to mean right, even though “left” and “right” are purely abstract concepts here.
Are you talking about time series graphs? They’re used to show how something has changed (gone up or down) over time. Time is conventionally always on the horizontal axis, because we’re used to thinking about going “forward in time” from left to right; and the thing that’s changing (profits, or gas prices, or squirrel population, or whatever) is on the vertical axis so that you can see the graph literally going up or down as the thing it shows increases or decreases when you move forward through time.
“That’s what facing mortality is all about: realizing that time is really the ultimate independent variable.” --one of my college math profs
Thank you all for all. Yes ThudlowBoink. That answers the question perfectly.
davidmich
Well just to clarify, the standard X-axis/Y-axis graphing system is called the Cartesian Coordinate System, named for the mathematician who invented it- René Descartes.
Thank you Hail Ants. Very much appreciated!
davidmich
Intuitively, one thinks of time as a horizontal metric, which does not change in value as it progresses, left to right, like words in a discourse. But the purpose of the graph is to represent the changes in some value, which is intuitively visualized as moving vertically. So a vertical representation on the graph represents something of changeable value, while the horizontal is for time, which is not a changeable value.
I once had the opportunity to speak (in a group) with some authority about graphs. This was a long time ago and I don’t remember who he was or what made him an authority about graphs. Maybe he had invented the abcissa or something.
Anyway, he said having the dependent variable vertical and the independent horizontal was more than a convention, it was a rule.
I just wanted to second this. This is a staggeringly beautiful book on the presentation of data – I never really appreciated that there CAN be nuance in graphs before reading this. If you’re at all into infographics and the such, it’s a must-read.
Also check out reddit’s section dataisbeautiful for a collection of nice infographics from around the web.
In information graphics, I think the usual term is fever chart for a chart that shows a trend line.
A *bar chart *shows discrete quantities, for successive years or different cities, without trying to connect the numbers into a trendline.
Of course, traditional terms are meaningless compared to whatever Microsoft Excel decides to call that type of chart.
Tufte’s books inspire, but I think the most insightful recent writings came from Richard Saul Wurman, such as his book Information Anxiety.
There’s really no difference between conventions and rules in mathematics, when you get right down to it. Maybe someday we’ll meet aliens, and discover that they put information into two-dimensional graphs like we do, except they prefer to put the independent axis vertical and the dependent axis horizontal. Or maybe they prefer to tilt the whole thing 45 degrees, or make the numbers increase going down the page, or to the left. Neither of those is really “wrong”.
I’ll order
Edward Tufte’s "The Visual Display of Quantitative Information.
and Richard Saul Wurman, such as his book Information Anxiety.
Thank you all. Very helpful
davidmich