Keywords: bump chart, time series analysis, change over time, rank.

Bump charts are useful for displaying change over time, for example in the volumes of topics or tokens. A benefit of bump charts compared to line charts is that they display both volume and rank during each time period, making it easy to see how the relative positions change over time.

Step-by-step guide

1. Add a Bump Chart workbench

To create a bump chart, click "Add workbench" and click the Bump Chart symbol. A Bump Chart workbench is now added to the workspace.

2. Specify the data to be used as streams

Drag the field with the data you want to display as streams, or the specific values that you want to display, to the "Group by" drop zone.

3. Specify the date field

Drag the field with the time units that you want to track change across to the "Dates" drop zone.

☝️ Note: Make sure the field containing time units are of "date" type and has the format you want to use (depending on whether you want to display years, months, dates, or something else on the time axis). If needed, use the Change Field Type operation and the Change Date operation to change.

4. Specify aggregation function

Define the aggregation function by dragging the desired field to the "Apply function & display chart" drop zone and select the function in the drop-down menu.

👉 Example. Consider a case where your dataset contains articles, each of which has an author, and where the articles have been split into segments, which then have been used to detect topics. There are now multiple interpretations of "topic volume". By applying the "count" function on the "id" field, the number of articles that each topic is associated with is calculated. Applying, instead, "count" on "segment.id" yields the number of segments in which each topic appears. Applying "count" on the field containing author names calculates the number authors that have articles associates with each topic. Finally, applying the function "sum" on the "topic.strength" field calculates the total strength of each topic, which is a way of accounting for the fact that topics can be more or less associated with a text.

The bump chart will now be displayed.

5. Adjust the settings if needed

The following settings are available:

  • Click the "Normalize" icon in the workbench header to normalize by time unit.

  • Click the "Show labels" icon in the workbench header to show the label of each stream in the graph.

  • Click the "Settings" icon in the workbench header to change the grouping source and select grouping values, date source and select date values, and aggregation function.

  • Click the "More" icon in the workbench header, then "Reset zoom" to reset the zoom level.

  • Click the "More" icon in the workbench header, then "Shuffle colors" to shuffle to colors of the streams.

Did this answer your question?