This article describes Crunch’s interactive time plots (graphs), which involves the following:
- Time plots require that you work with a categorical date variable as your columns/groups variable.
- Smoothing (also known as moving averages or rolling averages) can be applied to time plots, and also to tables where columns are categories with dates.
Making a time plot
To make a time plot, start by making a crosstab in Tables and Graphs mode using a categorical date variable as the column variable. The Display Controller (the grey panel that appears at the bottom of the screen) will then give you the option to show the data as a time plot:
In the following image, the categorical variable Favourite brand is in the rows and the categorical date variable Month is in the columns:
Location of details and labels
Crunch displays each category as a time series represented as a line. At the end of the line (series), the plot displays its value and label in the same color as the line. You can drag the wave title at the top to earlier waves which enables you to inspect the values and ranking of categories at a particular point in time.
In the following example, we’ve dragged from the end of the series in September back to July of 2017. This reveals the point at which Jeep at 19% overtook rival Honda:
The * asterisk button in the Display Controller allows you to turn on the confidence bands around the lines. These confidence bands show the margin of error, which is a 95% confidence interval around percentages.
In the following image, we’ve selected two brands, Ford and Honda, to illustrate:
The confidence intervals overlap around April 2017, which is suggestive that the difference between the points may not be significant (it is not a test of the significance of the difference). At all other points in time, the 95% error bands on the categories do not overlap, so it is likely that the difference is statistically significant.
In other words, in July 2017, Ford (39%) is statistically higher than Honda (18%). However, in April 2017, Ford (37%) was not statistically different from Honda (24%):
When plotting a variable with more than 5 categories, Crunch will start by showing you the top 5 categories, determined by their value at the most recent time period (i.e., the final column of the table).
Graphs with many lines plotted together are harder to read, and typically the user is most interested in the highest values, so this initial selection can make it easier to quickly make sense of your data. However, you can select which categories you want to display using the Select categories button on the right, and if you then save this analysis to the deck then this set of categories will be reflected (a) if you reload the analysis from the deck, (b) when exporting to PowerPoint, and (c) when viewing that analysis as a tile in a dashboard:
Having this control over what to show/hide can be very useful if you have a lot of categories, which could represent brands, statements, and anything else. Perhaps there is a category such as “Don’t know” you want to remove or perhaps your viewer wants to narrow the picture to only 2 brands (as we did in the above example for Ford and Honda).
The Select categories button allows you to have many categories available, but selectively hidden on demand.
Hovering your cursor over the lines will reveal the value at each point (and also the margin of error, if currently enabled).
You can turn on all the values simultaneously using the display options in the display controller (feature coming in March 2021).
Frequently asked questions
Can I customize the colors?
- The time plot, like any Crunch graph, will use the colors set by the dataset editor when viewed in a dashboard or exported to PowerPoint.
- You can edit the properties of a deck slide and edit the properties of a dashboard slide, to customize the colors for display and export.
Can I modify the x-axis labels?
- The labels of the x-axis are those of the underlying categories at the position of the date assigned to them. Editable labels are one of the advantages of using the new categorical date variables rather than date-time variables.
- You cannot modify the x-axis labels in the saved deck slide properties or the dashboard tile properties.
Can I modify the category labels?
- Yes, you can modify the categories in the saved deck slide properties or the dashboard tile properties. This only affects the display and export; it does not change the original row variable.
- Alternatively, you can change the labels in the original row variable.
Can I change the confidence interval around the lines?
No. The margin of error of a percentage is a 95% confidence interval around the point estimate.
Why am I not seeing all the categories?
When there are more than five categories, Crunch will start by showing you the top five categories, determined by their value at the most recent time period (i.e., the final column of the table). Graphs with many lines plotted together are harder to read, and typically the user is most interested in the highest values, so this initial selection can make it easier to make sense of your data quickly. However, you can select which categories you want to display using the Select categories button on the right, and if you then save this analysis to the deck then this set of categories will be reflected (a) if you reload the analysis from the deck, (b) when exporting to PowerPoint, and (c) when viewing that analysis as a tile in a dashboard.
Why do my line graph analyses not fill the space of dashboard tiles?
By default, Crunch’s line graphs are given a fixed aspect ratio designed to most appropriately display change-over-time data. If you would prefer your dashboard analyses to expand to fill the tile area, you can turn off this fixed aspect ratio by choosing “Fit to tile” in the dashboard tile’s properties.
Do the Crunch time plots export to PowerPoint?
- Yes—they export as editable line charts in PowerPoint.
- However, as PowerPoint has no concept of interactivity, the select categories button choices or the time marker will not transfer into the PowerPoint chart. You will get just the categories that you chose to display when you saved the analysis to the deck (although you can tweak this selection after saving by editing the saved analysis). Error bands are not a supported feature in PowerPoint, so cannot be shown.