Time-Series Data : Exploring data with time-series view
The time-based Keshif record view is the best way to see the time-series trends of all of your data at once. You can highlight, filter, compare across multiple selections (such as countries by region), get rich tooltips on mouse-over, and filter directly on the time-series charts.
This view allows three alternative plot types, available with a single click:
- Plot-by-value: to view data values as-is.
- Plot-by-ranking (aka. Bump charts): to view rankings and ranking trends.
- Plot-by-%change (aka. Time Indexed Charts): to compare time trends against a selected time.
All of these rich and intuitive features for data understanding and exploration are available automatically, with no effort, with all the time-series attributes in your dataset!
Opening Time-Series Record View
Click time-series button to switch the record display to time-series view.
Note: This view option is available only when your dataset includes a time-series data.
Highlighting and Filtering by Record Groups (Aggregates)
Highlight records by mouse-over on a record group (aggregate), such as countries by an income level.
Filter to the selected a group of records by a clicking on the aggregate.
Compare the selected group of records with another group by locking.
You can use any aggregation in the dashboard for highlighting, filtering, and comparing different groups of records. Just like in all other Keshif views!
The active mouse-mode defines your interaction options with the information-rich time-series charts.
Mode: Select Nearest Point & Record
In this mode, Keshif selects and highlights the closest data point & record to your mouse.
The selected record stands out with all of its values, and a tooltip shows the time & value of its closest point.
Click on the selected record to see all the details relating to the selected record (such as a country).
Mode: Select Nearest Time Key
In this mode, Keshif picks the closest time-key, and show tooltips revealing record values on that time-key.
This mode is useful to very quickly select and review an individual time key (such as year).
In this mode, you can filter records directly based on their time-series values.
You can even use multiple filters to explore records which improved or declined over time!
Note: This mode is only available when you plot by "value".
Plot Types for Richer Analysis
You can shift your focus easily between "value", "rank" and "change" plot types by just clicking the icons below the timeline. No other settings or configurations are needed!
Plot by Ranking (Bump Chart)
With the plot-by-ranking option, you can explore how record rankings change over time. This chart is a very powerful method to move beyond individual values, and see how records (such as countries) compare to each other, effectively enriching your understanding and analysis of data. For example, you can see how rankings of countries in a specific region change over time.
Plot by %-Change compared to a Fixed Time
With the plot-by-change option, you can explore how the data changes relatively over a fixed time. The values of all records at this time is set to 0% (baseline), and the records extend before and after compared to this individual baseline. You can choose this baseline time in the select menu next to the plot type button after you enable this view. The tooltips also show the relative change of the selected record(s) compared to this fixed time, giving you the context you need to better understand your data.
This plot type is commonly used in analyze stock market changes (how much a set of stocks increased/decreased over time), since the profit/loss depends on value change compared to a specific point. Unlike with Excel, QlikView, PowerBI, Tableau, or Vega's manual approaches, you don't need to enter any complex formulation in Keshif, and you can easily switch the compared time and across different metrics in a second. This chart is a great alternative view to keep in mind, beyond financial analytics.
Note: This feature is available only if your time series data has no negative values, so that the changes can be analyzed in a meaningful way.