6.1 Dynamic Scatterplot Chart Options

Figure 6.2 shows how BigML displays the dynamic scatterplot view. This view has two parts: a graph in the center of the image, and a data inspector on the right hand side.

1. The graph in the center of the image is the visualization itself. You can configure the graph options highlighted in Figure 6.2 using:

   • Fields selectors: you can select two fields from your dataset, one field for each axis (Y and X). Fields must be either categorical or numeric (text, items and date-time fields are not supported as scatterplot axes).

   • Logarithmic scale: either axis with numeric fields may be plotted logarithmically, useful when your values have a large range.

   • Regression line: you can show and hide a regression line when the two selected fields are numeric. The regression line fits a simple linear regression to your data, useful for highlighting trends.

   • Create a dataset: you can select an area in the chart (by clicking and dragging in the chart surface) and create a new dataset containing only the data points in the selected area.

   • Get new sample: when your dataset is very large, BigML automatically takes a random sample of 500 instances so you can better visualize the data points in the chart. With this option you can visualize a new sample of your dataset.

   • Export chart you can export your chart as an image (PNG) with or without the legend.

   • Freeze the current view by mousing over the data point you are interested in and pressing shift on your keyboard. You can release the view by pressing the esc on your keyboard.

   • Zoom: you can zoom in by selecting the area in the chart that you want. You can zoom out again by clicking anywhere within the chart area.

   • For your numeric fields, BigML also computes the Pearson and Spearman coefficients so you get a measure of the linear correlation between your chosen fields.

   • Color selector: you can select a field to color your chart points (text, items and date-time fields are not supported).

   

2. Finally, the data inspector on the right hand side shows all your dataset fields, distributions, and values (see Figure 6.4 ). When you mouse over a data point in the chart, you will see the values for each field highlighted in the corresponding histograms. You can freeze this data point view by pressing the shift key. To release it, just press esc on your keyboard.