Personalize the feel and presentation of your data with a deep dive into customization options for the appearance of your graphs.
You will learn how to:
Personalize the feel and presentation of your data with a deep dive into customization options for the appearance of your graphs.
You will learn how to:
This video is part of the Graphing Basics series, presented by Dr. Trajen Head, Product Manager for GraphPad Prism.
Transcript:
So in the first video of this series we looked at some simple and direct ways that we could use to add some customization and personalization to our basic default graph. In this video we're going to continue looking at graph customization, focusing on the options found within the format graph dialogue.
To access the format graph dialog, you can either click on the format graph button in the toolbar or simply double click in the graphing area itself. The dialogue that appears has a lot of options and may seem a bit overwhelming at first, but the choices you can make in this dialogue are all relatively intuitive and will start to become very familiar as we work through various options available and you begin customizing graphs of your own.
We'll start on the appearance tab of this dialog and right at the very top you can see that there is a dropdown that allows you to select which dataset you'd like to customize. You can choose an individual dataset to modify or you can select to change all datasets at once.
As a shortcut to apply customizations to all of the datasets at once, you can hold the control button down on your keyboard, which will cause a small globe to display next to your cursor. This indicates that you're making changes to all of the datasets.
The next section of this tab provides available options for the graph style depending on the family and type of graph you've selected. Under appearance, you can select how you want the data to be shown. These options are similar to the option that we saw when choosing our graph type and the previous video. For example, you can choose to show only a single point for the mean or median, you can choose to show mean or median with error bars or you can choose to show each replicate. If you choose each replicate, as we've done here, you can choose to have the points of the graph aligned or staggered using the options under plot. Alternatively, if you choose mean or median with error, you can use plot to identify how you want the air to be displayed of standard deviation, a standard error of the mean, the 95% confidence interval or the range of the data.
The order and presence of the objects that can be customized in the next section may vary depending on the family and type of graph that you're customizing. However, for each graph, this section will provide appropriate options for displaying symbols, bars, spikes and drop lines, error bars, connecting lines and curves, area fill on the graph and other options such as bars and boxes, which is not shown here as our graph type doesn't include any of these objects.
Once again, depending on the graph type being used and the options chosen under appearance and plot, you can choose to toggle some of these object sections on or off while others may be disabled entirely. For example, we currently have mean in error selected under appearance near the top of this tab. Because of this, we have the option of customizing the appearance settings for the error bars on the graph or the option to toggle the error bars on or off entirely. If we switch our selection back to each replicate under appearance, you can see that the options for error bars have simply been disabled.
In the symbol section of this dialogue, we can choose to change the color, shape, size, and border options for the symbol shown on the graph. If I wanted to change only the symbols for Group A, I could use the dropdown menus to pick a new symbol, a new size, and then click apply. You'll notice that this only changed the dataset that we have selected at the top, dataset A in this case. If we either hold down the control key on our keyboard, or choose change all datasets, we can choose to change that assemble shape, size, and border thickness for both datasets. Note that as long as we don't manually select an option while making global changes, each dataset will retain its own setting. This is why in order to change all datasets, I needed to reselect the symbol, shape and size. However, as long as I don't manually select a color, each dataset will keep its own color. If we were to change the color while change all datasets was selected, however, that color will be applied to all of the datasets as you would expect.
Let's make a few more changes to our graph. Let's choose to show the connecting lines, change its thickness and pattern, set it to connect the means instead of each replicate and choose to leave a gap at symbols. Let's also choose to show the area fill. We want the fill to show below the line and although we could change the colors, we would likely want to do that individually for each dataset, not globally. For now, let's just click apply.
Immediately we see a problem. Because our first dataset has values that are higher on average than our second dataset, we can't see the majority of the area fill for the second dataset. It's covered by the first. One way around this could be through the use of semi-transparent colors for the area fills.
Let's select each dataset individually and adjust the transparency of the fill color. You can now see that both area fills can be seen when we click apply. However, some of the pink symbols appear darker as they are behind the blue semi-transparent area fill. To address this, we can utilize another useful method for arranging the order of datasets on a graph. On the format graph dialogue, we can click on the second tab at the top, datasets on graph.
The first thing you'll see on this tab is the large section that will list each of the datasets shown on the graph. In this window, you can click and drag to reorder the datasets however you'd like. For example, we can drag dataset A down so that it is below dataset B. When we click, we can now see that the pink symbols are in front of the blue area fill and must appear to be the correct color again.
Other useful tools on this tab include the ability to add, replace, or remove datasets on the graph. In fact, you can add or replace datasets from other data tables in the same Prism file, allowing you to graph multiple sets of data together without having to copy of the data into a single data table.
With a large number of datasets, it's also sometimes easier to use the reorder buttons to organize the order of the datasets on the graph. You can choose to send a dataset to the top or the bottom of the list, move it up or down one position, or reverse the order of the entire list of datasets at once.
Finally, on this tab, there are additional options that can be used to modify datasets individually. Depending on the graph type, these options will be somewhat different. For example, for an XY graph, you can choose to nudge the data in the direction of a specified number of X and Y units. This can be useful when data from multiple datasets perfectly overlap. I would, however, caution against using large values when nudging your data, as this obviously results in the data appearing at a new location and could skew it's interpretation.
For datasets with a very large amount of data, you can also choose to have Prism graph one row of data, then skip a defined number of rows. This choice may help speed up graphing, but it comes at the obvious cost of excluding data on the graph.
Let's look at a new tutorial dataset for a new graph type to explore the remaining options on the format graph dialogue. We'll close this dialog and create a new data table. We'll select group, ensure that start with sampled data to follow tutorial selected and then choose ordinary three datasets under two way anova and click create.
As before, we'll simply go straight to the graph. In the change graph dialogue, you can see that our options are a bit more extensive for this family of graphs than they were for our first example, but the concept is exactly the same. Using the tabs, you can choose to show individual values as a scatterplot, scatterplot with bars and other options. You can choose the box and violin tab to share the data as box and whisker plots or as a violin plot.
We're going to choose summary data so that we can create a simple bar graph. We can choose to start with interleaved, separated or stack bars. Let's just choose interleaved and click okay. Just as I showed in the previous video, you can adjust the size of the graph, change the fonts, or apply a color schemes as you'd like. Let's change the color scheme to colors to more easily distinguish between the groups of bars on our graph before we look at some of the other features in the format graph dialogue.
To open the format graph dialogue, we could double click on the graphing areas before or click on the format graph button in the toolbar. This time when the format graft dialogue opens on the appearance tab, you may notice that the specific available have changed slightly. We have different options under the style appearance dropdown and the objects on the graph that you can customize are different, now including bars and boxes. However, the concept is exactly the same and you can customize these using the same process we looked at previously.
If we click on the datasets on graph tab, we still have the ability to rearrange the datasets as before, but we also now have some new arrangement options for this family of graphs at the bottom of the dialogue. These options all deal with how we want Prism to display each dataset on the graph in relation to the dataset before it. That may sound confusing at first, so let's look into the couple of examples to get you more comfortable with this concept.
Obviously, if we have the first dataset selected, there's no previous dataset or dataset before it, so we can't make adjustments in these options. So let's actually select the last dataset. Now we see that these options are available to modify. Originally we had selected to show an interleaved bar graph and that's reflected here, with each dataset interleaved with the one before. However, we could choose to change dataset C to be stacked or super imposed with dataset B.
We saw in our previous example that the order of the datasets in this window determine the order of which datasets are shown in front or on top of the others. Note that if superimposed is selected for bar graphs, tourism will automatically show the smaller Dataset in front regardless of the data order so that a smaller bar isn't completely hidden behind the larger bar.
Finally, in addition to interleaved, stacked and superimposed, we can also choose to have dataset C separated with respected to dataset B. When we click apply, we see that while dataset B is still interleaved with dataset A, dataset C is now separated off by itself. We can further emphasize the separation using the last tools on this tab, including the ability to add a vertical line between the selected dataset and the previous one, or adding an additional gap between these datasets.
The next tab of the format graph dialog contains a large number of options for graph settings, which, again, are going to be largely dependent on the family of grafts that you're using. In this case, we can choose to show the graph in either the horizontal or the vertical orientation. You can also specify where the baseline occurs for our bars. For example, we may wish to have the bars begin at Y equals 50. We can see that this value is actually larger than some of the bars on the graph, so when we click apply, those bars now go down instead of up from the baseline. We can also choose to either show or hide the baseline from the graph. Let's hide it.
The next section provides a number of options for setting the spacing of the graph. You can control the amount of space, if any, that Prism leaves for blank data in the data table, the amount of space between bars of a single group, additional space between groups of bars and the amount of space before the first and after the last bar on a graph. For example, let's change the space between adjacent data to 0% and click apply.
These are all very small ways that you can personalize the feel of your graphs without dramatically changing their presentation. Note that you can also choose to format individual bars or points without changing the entire dataset by right clicking or control clicking on the desired bar and choosing the option format this bar. If you choose to apply an individual formatting to one or more bars, you choose to remove individual format in here in the format graph dialogue and revert all bars or points back to the formatting of their dataset.
Finally on this tab, a very minor feature, the one that many people really like is the ability to choose if the legend icons for the bars are displayed as a rectangle or square.
The last tab of the format draft dialogue is the annotations tab and is unique to bar graphs. This set of options provides yet another way that you can customize these graphs. We can turn on imitations on our graph by checking the box beside write values in or above the bars, which enables numerous other options for customization. You can choose which values you want shown, if you want the text oriented horizontally or vertically, the position of the annotation, either above the bar and error bar, within the bar at the top, well within the bar at the bottom of the bar, and options for formatting the value being displayed.
Let's say I choose to show the plotted value of the mean in this case with horizontal text above the bar and error bar, in decimal format with one decimal place. When I click okay, you can now see that the graph has been automatically labeled with the appropriate values. I should note here that if the data for any of these bars is changes in the data table, both the value displayed and the position of the annotation will change automatically.
In the next video, we'll look at one more important set of options for customizing your graph. Those options found in the format axes dialogue. We'll keep using this same graph so you can either continue now or save this graph to explore those options later. Thanks for watching.