This document will help you identify important considerations when using graphs (also called charts) to represent your data. First, it is crucial to understand how to create good graphs. Then, an overview of different types of commonly used graphs is provided. Important Considerations For Graphical Representations Of Data Identify the message you are trying to get across with the graph o make sure your graph speaks the message clearly o identify trends or differences in the data o be accurate in a visual sense Use appropriate graph types (and use them appropriately) o show data without changing the data s message make sure it s representative o emphasize appropriate elements Keep the graph (and it s message) simple, clear, and visually appealing o you should be able to tell what it is generally about just by glancing at it o avoid clutter too many items, too many colors, too many textures = bad design o take into consideration how the data will be presented will it be readable? (will it be printed in color or black/white, printed full-size or small-scale, projected?) o if you can t seem to limit the amount of items on a graph, consider breaking the elements down into multiple, more simplified graphs Be consistent o use similar graph types for graphs that convey the same type of information or use the same type of data o use similar color schemes for graphs that convey similar information o use consistent sizes and styles of text Avoid ambiguity o label data elements clearly to avoid confusion o don t use colors or textures that are too similar o make sure your text and background colors work well together o avoid text that runs together or is illegible Example on left (poor example) is too cluttered, colors that are too similar have been used for precipitation figures, and the text is not easy to read. The gridlines are misleading because the graph has two y-axes. The example on the right (better example) has more definition in the colors, used different data indicator (line) for the acres consumed since it is plotted on a separate axis, and added an image to the background that helps the message of the graph: that the driest months are more prone to fires.
A Note about Scale: With all chart types that depict trends in data over time, it is very important to use the correct scale. Incorrect (or poorly chosen) scale can give an incorrect impression about the data. Consider the following example: the chart on the left uses a y-scale from 93 to 101, and the difference between the data points is much more pronounced than the chart on the left, whose y- scale is from 1 to 101. Percent of Data Completeness Percent of Data Completeness 101.0 101.0 100.0 91.0 81.0 99.0 71.0 98.0 61.0 97.0 51.0 96.0 95.0 41.0 31.0 21.0 94.0 11.0 93.0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 If you are attempting to show that your data are very complete, you might want to use the chart on the right. Using the adjusted scale like the chart on the left shows more of a difference between data points, and would be more useful for showing the difference in PM concentration or temperature). 1.0 Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8 Day 9 Day 10 To adjust the scale of a graph, you can select the axis you want to adjust and then double-click it. The Format Axis dialog box will appear; select the Scale tab. If a checkbox on the left-hand side is checked, it means that Excel will automatically set the scale values for you. If you want to set the values yourself, click the text box to the right of the label and enter the desired values. The minimum value is the lowest value displayed on the axis, and maximum denotes the highest value displayed (typically several units higher than your highest data point). Major unit indicates where a major gridline will be displayed, and minor unit where a minor gridline is displayed. Whether or not the major or minor gridlines appear on your chart depend on the options set in the Chart Options box.
INTRODUCTORY GRAPH TYPES Column and Bar Charts Bar and column charts show data changes over a period of time or illustrate comparisons among items. Values in bar charts are arranged horizontally, while values in column charts are arranged vertically. Important considerations with Column and Bar Charts: Column and Bar charts are best suited for data with fewer than ten (10) data points (bars) The greater the length of the bar, the greater the value Column graphs lack much room for a written label at the foot of each bar; so it is best to use a column graph when the label is short. Bar Charts can display longer labels for each bar *Simple or Clustered This type of chart compares values across categories. It is also available with a 3-D visual effect. As shown in the following chart, categories can be organized horizontally, and values vertically, to emphasize variation over time. Clustered charts depict data from multiple categories (no more than three) for each data point. This is an easy-to-read chart depicting temperature highs and lows for one week. Comparisons can be made across single categories (High Temp or Low Temp), or between categories (compare High and Low Temps for each day). *Stacked This type of chart shows the relationship of individual items to the whole, comparing the contribution of each value to a total across categories. 100% Stacked charts compare the percentage each value contributes to a total across categories. It is also possible to use a 3-D visual effect for stacked charts. This is an example of a 100% stacked bar chart (with 3-D effect) depicting PM concentrations (two different types) for one week. The total of the two concentrations equals 100%, therefore the chart is depicting percentages instead of actual values.
Pictographs These chart types use pictures to represent data similar to column or bar charts, but can be more effective at representing the subject matter of the graph (using an appropriate image). Data can be represented with a single image stretched the length of the column or bar, or images can be stacked to represent the data in a scaled proportion. Care must be taken in using pictographs because the image size may not always be representative of the true data (sometimes this is actually why they are used). In Excel, images can be stacked according to the size of the selected image, or the size can be scaled to be proportionate to a selected value (in the picture to the right below, each image represents 12 families, but if the scale is not specified, one must be skeptical of the representativeness). Line Charts A line chart shows trends in data over time. Line charts represent data in a similar fashion to Column Charts, and have many of the same sub-types. However, Line Charts are best used when there are numerous data points for a category and/or multiple categories to compare, as they are less cluttered than Column Charts. Line charts can be made with or without the data markers ( dots on the lines). This is an example of a line chart with two categories (PM10 concentration and Temperature) for one month of daily data. It is easy to compare the two categories and change over time. Note, however than since the two datasets have different units, you are only able to compare how they relate to each other (i.e. when deg. F goes down, PM concentration goes up). Data can also be plotted on separate axes.
Pie Charts A pie chart is best for showing parts of a whole. It shows only one data series and is useful when you want to emphasize a significant element in the data, or percentages (pieces) of a whole. However, one disadvantage is that it can be very difficult to see the difference in slice sizes when their values are similar. This is why it is important to label the slices with actual values. Pie Charts can be made in standard format, exploded, and/or with a 3-D effect. Additionally, Pie Charts have special types that show a breakdown of an individual slice using a second Pie or a Bar Chart (called Pie-of-Pie or Bar-of-Pie, respectively). See The Power of Graphing (Part II) for more information. This is an example of an exploded pie chart where the pieces are separated from the circle. Notice the data labels on this chart denote the percentage each piece represents of the whole. Another option would be to use actual data values rather than percentages. Area charts An area chart emphasizes the magnitude of change over time. Area charts are line charts with the area beneath the lines filled in with colors or patterns to highlight the importance of one or more categories. Like Bar, Column, and Line charts, Area Charts are available in 3-D effect, stacked, and 100% stacked sub-types. It is important to be careful with series order when creating area charts, especially stacked area charts (see examples below) *Stacked Area This type of chart displays the trend of values over time or categories, with an emphasis on the sum of the plotted values. A stacked area chart also shows the relationship of parts to a whole. A simple area chart displays only true data points rather than totals.
This chart is a Stacked Area Chart, which depicts particulate concentration data from 14 days of monitoring both PM10 and PM2.5. Notice how in this chart (top left) it appears that the PM10 values reached nearly 50 µg/m 3 on Day 9. The stacked area chart depicts the data as parts of a whole and is combining PM2.5 & PM10 into what could be called a Total Suspended Particulate (TSP) concentration. The Area Chart (bottom left) depicts both data series independently, and it can now be interpreted as showing how the two datasets related to each other. To change a Stacked Area Chart to an Area Chart, go to Chart Type and select the first option. If all series do not appear as desired, doubleclick the series in the chart itself and select the Order tab. You can then move each series up or down to get the desired image.