Time series graphs can be used to visualize trends in counts or numerical values over time. Because date and time information is continuous categorical data (expressed as a range of values), points are plotted along the x-axis and connected by a continuous line. Missing data is displayed with a dashed line.
Time series graphs can answer questions about your data, such as: How does the trend change over time?
A crime analyst is studying crime trends in his city to determine if initiatives to reduce crime have been effective. The analyst uses a time series graph and compares the number of incidents over time to the timing of the crime reduction programs.
The graph above shows peaks in incidents in the period from May, 2014 to October, 2014 and again from May, 2015 to September, 2015. There is a decrease in incidents in February, 2014 and 2015. If the crime reduction programs occurred in those months, that would imply that the programs were effective, but only for a limited time.
The analyst realizes that looking at a total number of incidents over the two years only tells part of the story. Subgrouping the time series graph by another field, such as the crime type, can give more insight into the effectiveness of the programs for different types of incidents.
The grouped time series graph above indicates that the majority of incidents were categorized as Larceny. The counts of larceny show a similar pattern to the ungrouped time series. The other incident types, burglary, motor vehicle theft, and robbery, have relatively stable counts across the two year time period, with the exception of a spike in burglaries in May, 2015. Based on what the analyst sees in the grouped time series graph, he will recommend focusing the programming on reducing the number of larceny crimes in the city.
The following cards may be helpful to pair with your grouped time series graph:
- A related map styled with the same field you used to group your time series graph, such as Day Of Week. When you interact with the time series graph or the map, you'll be able to see simultaneous temporal, categorical, and spatial patterns.
- A related bar chart using the field you used to group your time series graph to see highest and lowest values by your chart subgroup.
Create a time series graph
To create a time series graph, complete the following steps:
- Select one of the following data options:
- One or more date/time fields
- One or more date/time fields plus one or more number or rate/ratio fields
- One date/time field plus a string field
- One date/time field plus a number or rate/ratio field plus a string field
If you do not select a number or rate/ratio field, your data will be aggregated and a count will be displayed.
You can search for fields using the search bar in the data pane.
- Create the time series graph using the following steps:
- Drag the selected fields to a new card.
- Drop the selected fields on the Time Series drop zone.
You can also create charts using the Chart menu above the data pane or the Visualization type button on an existing card. For the Chart menu, only charts that are compatible with your data selection will be enabled. For the Visualization type menu, only compatible visualizations (including maps, charts, or tables) will be displayed.
Time series graphs can also be created using Time Series, which is accessed from the Action button under Find answers > How has it changed?
When you drag a date/time field onto your page, the Chart drop zone is replaced with the Time Series drop zone. The reason for the change is because a time series graph is the only chart type that can be created using a date/time field.
The Legend button opens the Layer options pane. The Layer options pane contains the following functions:
- The Legend tab is used to view the symbols on the chart. The pop out legend button displays the legend as a separate card on your page. You can use the legend to make selections on the chart. To change the color associated with a value, click the symbol and choose a color from the palette or enter a hex value (unique symbols only).
- The Options tab is used to show or remove the time slider from the card or show a gradient fill below the time series line. If the time slider is removed, you can still zoom to a time interval by clicking and dragging your pointer across the chart.
- The Style tab changes the symbol color (single symbol only) and the pattern and thickness of the line.
The trends for the time series graph can be symbolized as a count of features over time or as a number or rate/ratio field. The values in the trend line can be calculated as the sum, minimum, maximum, average, percentile, or median of values for the chart.
The x-axis shows the date/time fields visualized on the chart. The arrow next to the field names can be used to select or unselect date/time fields to display on the chart. If multiple date/time fields are selected, the y-axis can only display one number field or the count.
The y-axis shows the count or number fields and statistics visualized on the chart. The arrow next to y-axis can be used to select or unselect number or rate/ratio fields to display on the chart. If multiple number or rate/ratio fields are selected, the x-axis can only display one date/time field.
Drag additional date/time fields or number fields to an existing time series to add them to the chart. The fields must be from the same dataset used to create the chart.
You can optionally select a Group by field on the x-axis. The Group by field must be a string field and is used to divide the trend line into subcategories. A Group by field is only available on time series graphs containing a single date/time field and a single number field or count.
Use the Chart statistics button to display the mean, median, upper quartile, lower quartile, or a custom value.
Use the upper and lower time slider handles to zoom in on a specific time interval. You can also drag your pointer on the chart to zoom to a specific interval. Use the Default extent button to zoom back to the full extent of the data.
Zooming to a time interval will not select or filter your data or change the binning intervals.
Click on the y-axis to switch the scale between Linear and Log.