Charts are Einblick's primary visualization tool. You can use them to display key insights, quickly profile or filter your data, and much more.
Creating a Chart
Charts can be created in two ways, either by dragging out a attribute onto the canvas, or by creating it the same way as other operators.
Dragging attribute onto canvas
You can drag out an attribute using either the left menu after expanding a dataframe, or from a table column header. When a chart is dragged out in this method, a bar chart with the attribute count distribution is automatically populated.
Configuring a Chart
The primary way to configure a chart is by changing the
Chart type, this will automatically configure your chosen axis attributes so that you get the chart you expect. If you require more advanced control over chart rendering, the
Pivot chart chart type will remove any restrictions on the configuration settings.
Chart types. From top-left, going clockwise: Bar chart, Vertical bar chart, Line chart, Area chart, Scatter plot, Heat map
The remaining option menus all have the following options:
- attribute: the column in the table to act on
- binning: whether to bin the selected attribute based on the corresponding property. Options include:
- No binning
- Equi-Width binning: Uses bins of equal size
- Natural binning: Creates bins by using k-means clustering along a single dimension
- aggregation: the function used to determine the value of individual marks (points, bars, etc.) Options include:
- No aggregation
- distinct count
- standard deviation
- sort: the sort order of data points
- include zero: whether to include
0in the corresponding axis or scale
Axes: x-axis, y-axis
Using these menus, you can choose what attributes to include on each axis. In most cases, you must select two axes for the chart to render.
To provide additional information in a chart, coloring can be added according to another attribute. The way the coloring affects a chart is dependent on the chart type.
Splitting chart on attribute using color
Bar chart, Vertical bar chart, Line chart
Scatter plots will also default to splitting attributes by color, but can also be configured to change the mark color.
Changing mark color
To configure a scatter plot to change mark color, you need to change the color attribute to an aggregation.
Changing color scheme
Similarly to color, mark size can be used to illustrate additional information. Currently size is only usable with Scatter plots.
Row and Column
The row and column options allow you to show multiple plots, each plotting a subset of data points, based on the values of another attribute. This is very useful for doing a breakdown of a existing chart by examining how the change in a attribute affects the data.
Mark type determines the kind of plot. The available options are
point. This option is only available if you select the
pivot chart type, and is automatically configured by the
Data Point Rendering Limit
There is a default maximum of 200 points that can be rendered in a chart. Therefore, it's recommended to use binning and/or aggregation in a chart in order to to reduce the number of points that need to be rendered. If some points are not being visualized, you will see a warning prompt next to the chart.
If binning or aggregation cannot be used, or if even after their application, more than 200 points need to be rendered, right-clicking a chart and selecting Adjust rows in view will give you the option to change the number of visualized points. Note that if too many points are plotted, performance issues could arise, so it is recommended to keep the number of rendered points relatively small.
Charts can be used as filters for other operators. To use a chart as a filter, drag the filter icon onto the input of another operator. Next, select the data points you want to include in the filter. The following example shows an example of selecting points from a chart to filter the values in a table.
Filter selections can also be made by using CTRL-Click and circling the desired selections.
Filters can be managed using the Active Selections Menu, clicking on the menu will give a list of current slections, and give you the chance to remove them individually or clear all of them.
To manage a filter link, you can click on it to bring up the Filter Link Menu. This menu will allow to invert the selection (by selecting
filter out vs
filter by), delete the filter link, and view active selections.
Similarly, rather than using selections as filters, they can be used to create visual highlights. Tapping a link/arrow will convert the filter link to a brush link, causing the selected subpopulation to be highlighted in the connected chart.
As with filters, an arbitrary number of brushes can be defined. Each brush gets its own assigned color. In case of an overlap of two brushes (e.g., leased cars that have not been used, in the example below), the overlap is indicated using an additional color.
When multiple filters are used, you have the option of using either AND or OR logic between the filters. Using AND will filter down to elements that satisfy all conditions, while using OR will filter down to elements that satisfy at least of the conditions. Using combinations of ANDs and ORs is currently not supported.
To toggle whether AND or OR logic is used, click on the filter icons leading into the targeted dataframe and select the
Show applied filters option. There should be a dropdown to change between
Charts can easily be exported to PowerPoint. Select
Export to PowerPoint in the chart menu and name your export appropriately to make the chart available in PowerPoint. To understand this in more detail, visit the PowerPoint Integration page.