Plot the points in the data set. Place each point at the intersection of the independent variable's position on the horizontal axis and the dependent variables position on the vertical axis. For the coconut example, rainfall would be along the horizontal axis and number of coconuts would be on the vertical axis.
Look for a pattern that implies some connection between the variables. The pattern will be evident in the spaces in the graph that are crowded with points and the places that have few or no points.
Identify outlying points and remove them if justified. If almost all the points are grouped together in some pattern, with just one or two lying outside the group, you might have reason to ignore the outlying points. For example, you could find that one of the coconut groves is right next to a river that supplies it with extra water. You could then justify throwing that point out because that grove is not typical.
Examine the graph for a linear relationship. If the points have tightly grouped themselves from upper left to lower right, for example, there is a strong inverse relationship between the variables: that is, as the independent variable gets larger, the dependent gets smaller. You can then use linear regression to determine the exact form of the relationship.
Look for a nonlinear pattern. Possibly there is a tight grouping headed from lower left up to a peak in the middle, then trending back down to the lower right. That would indicate the variables are related, but not with a simple linear relationship. A nonlinear regression analysis will give you the mathematical form of the relationship.
Look for evidence of randomness. If there is no pattern in the data, you may be forced to conclude that the relationship you suspected does not exist. You'll have to apply your expertise to see if there could be other factors obscuring the relationship. For example, the coconut groves may be growing in different soil types. Perhaps within each soil type the relationship between rainfall and production might be strong.