# Get Started

# Load a Dataset

Drag a Data Source

onto the canvas. Click on the created data source and its option panel will pop up on the right of the screen. Click in the option panel to select a dataset to load in the data source.

# Create a Visualization

Let's create a Scatterplot

to visualize the dataset we just loaded. After dragging a scatterplot to the canvas, build an Edge that connects the Output Port of the data source to the Input Port of the scatterplot. To create an edge, you may drag from the output port and drop it either at the scatterplot node or its input port.

You may change the X Column and Y Column and other settings of the scatterplot in its option panel.

# Forward Interactive Selection

Interactive selection in a visualization can be forwarded to another node for detailed exploration. We can create a Table

to inspect the selected points in the scatterplot. Connect the Selection Port of the scatterplot to the Input Port of the table. When the selection in the scatterplot changes, the newly selected rows are reactively shown in the table.

# Highlight Interactive Selection

VisFlow allows you to assign visual properties to data items so that interesting subsets of the input data can be brushed and linked across multiple visualizations. You may use a Visual Editor

to set the visual properties on the data items. In this example, let's highlight the selection from the scatterplot in a histogram.

First create a visual editor for the selection from the scatterplot. Set the Color visual property to red in the option panel of the visual editor.

Then create a Set Operator

and set its Mode to Union. Merge the highlighted selection into the full dataset by connecting both the Output Port of the visual editor and the Output Port of the scatterplot to the Multiple Input Port of the union node. Finally create a Histogram and connect it to the Output Port of the union node.

This flow diagram shows the distribution of the selected data points from the scatterplot in the histogram.

# Filter the Data

You may add an Attribute Filter

to find data items based on their attribute values. In this case let's find the cars that have an American origin.

Create an attribute filter and connect it to the Selection Port of the histogram. Keep the filter type as Pattern filter, and set the column to be origin in the option panel. Add a pattern American in the patterns list below.

Now draw a rectangular box to select a few stacked bars in the histogram. The attribute filter will find among the selected cars in the histogram those that have an American origin. To show the details of those cars, we may further add a Table

. Feel free to explore other filter types as you like. For more details on how to use an attribute filter, see the attribute filter documentation.

# Subset Flow

You may have noticed that the red color assigned earlier by the visual editor keeps in effect in the histogram and the table. This is because VisFlow uses a subset flow model that renders everything with respect to their visual properties.

In the subset flow, all data transmitted in the dataflow are subsets of table rows from an input table. The input table rows are never mutated, except that visual properties can be assigned and modified down the dataflow. Such a design ensures that visual properties can be used to keep track of subsets unambiguously, so that brushing and linking can be well supported for visual data exploration.

# Working with Multiple Tables

If you have more than one input table, you may relate them using table linking. A Linker

may be used to find table rows from one table using the attribute values extracted from the other table. Technically, a linker extracts the attribute values from one table as constants, and use those constants to filter the other table. Read more about the diagram elements definitions here.

# Next Steps

That's it! You have successfully completed the basics of VisFlow. Learn more about the dataflow model behind VisFlow in the Dataflow Section. Checkout the Node Types in VisFlow for the detailed usage of each node type.

Alternatively, you may go ahead and freely explore VisFlow yourself. Remember to come back whenever you have questions about specific aspects of the system.