- a way of exploring the relationships in data
- a way of displaying and reporting data, making it easier to report patterns and relationships, shapes of distributions, and trends.
Any graph used to report findings should show
- the significant features and findings of the investigation in a fair and easily read way
- the underlying structure of an investigation in terms of the relationships between and within the variables
- the units of measurement
- the number of readings (though sometimes these will be in the accompanying text)
- the range and interval of readings, where appropriate.
Original source: http://arb.nzcer.org.nz/supportmaterials/tables.php
Tips for Good Graphs
1. Give your graph a title. Something like "The dependence of (your dependent variable) on (your independent variable)."
2. The x-axis is your independent variable and the y-axis is your dependent variable.
3. LABEL your x-axis and y-axis. GIVE THE UNITS!!
4. When graphing data from lab, make line graphs because they tell you how one thing changes under the influence of some other variable.
5. NEVER connect the dots on your line graph.
Why? When you do an experiment, you always make mistakes. It's probably not a big mistake, and is frequently not something you have a lot of control over. However, when you do an experiment, many little things go wrong, and these little things add up. As a result, experimental data never makes a nice straight line. Instead, it makes a bunch of dots which kind of wiggle around a graph.
To show that you're a clever young scientist, your best bet is to show that you KNOW your data is sometimes lousy. You do this by making a line (or curve) which seems to follow the data as well as possible, without actually connecting the dots. Doing this shows the trend that the data suggests, without depending too much on the noise. As long as your line (or curve) does a pretty good job of following the data, you should be A-OK.
Original source: https://chemfiesta.org/2014/09/17/drawing-good-graphs/