# Explanation the difference between various types of plots

1. Oct 25, 2012

### Ein Krieger

Hello,

Why people use semi-log and log-log plots? What are they advantage?

Linear plots are more convenient as they show the direct relationship between x and y variables, aren't they?

Thanks.

2. Oct 25, 2012

### Staff: Mentor

There are probably lots of reasons that vary by technical field. In electrical engineering (EE), we use semi-log and log-log plots because our circuits operate over wide frequency ranges (many decades of frequency, typically), and the gain (frequency response) of the circuits can vary quite a bit as well. When you use a log plot, a 1/f relationship turns into a straight line, whose characteristics are much more obvious (like what is the slope?).

http://csserver.evansville.edu/~ric...bode_plots/html/bode_plots_html_m27115e71.png

.

3. Oct 25, 2012

### AlephZero

One reason for making logarithmic or semi-logarithmic plots is that the turn other types of functions into straight line graphs.

For eaxmple if $y = a^x$ where $a$ is an (unknown) constant, then $\log y = x \log a$ and you can find $a$ from the "straight line" graph of $\log y$ against $x$.

If $y = x^a$, then $\log y = a \log x$ and you can find $a$ from a graph of $\log y$ against $\log x$.

If you are measuring data, it's easier to "see" how close the points are to a straight line than trying to guess how well points lie on the "right" curve, especially if you don't know what the right curve actually is!

4. Oct 25, 2012

### lurflurf

Transformed plots are useful when they simplify the relationship of the data. Of course you need to know what transformations are helpful in advance, and be ready to recognize when things have gone wrong. Most of the time it is easier to recognize the transformed pattern than the natural for, but there is a danger of tricking your self into seeing a pattern wit ha transformation.

5. Oct 25, 2012

### symbolipoint

Some kinds of plots such as log-log or semi-log may more easily allow the graph to appear as a line. We like NEAT relationships that we can see.

In cases in which semilog or log-log do not give appearance of a line, some other type of data treatment might give one.