The scatter chart shows the relationship among two metrics, for instance between units and price, in the t0 and t1 periods.
Small datasets are plotted as scatterplots (that show a regression line, and the value of the regression coefficient).
Large datasets are plotted as binned heatmaps and data is plotted as rectangles. The darkness of each rectangle is proportional to the number of observations encountered.
The plot is run at the total plotted data level.
The total level scatter is accompanied by its Residual plot,...
...and by its QQ plot to check regression validity.
A plot is also run for the chosen dimension, for small ....
...and large datasets
One dimension is plotted at a time. You can choose which dimension to plot with this widget.
SAVING THE PLOT
To save the plot as a PNG file, click on the small circle to the right of the legend, and select "save as PNG".
Open the expander under the 📊 Plot type drop down menu by clicking on the ➕ sign.
Additional options are available for selection.
The Y-axis metric and the X-axis metric widget allow you to change the axes of your plot.
You can change the number of top items, and you can choose not to plot the value of the other aggregated items.
Values can be plotted in absolute value, as a % of total dataset, as a % of total dataset after filters, or as a % of the result row. The last two options only make sense if you plot a result row.
Choose whether to exclude outliers from the plot. If this widget is set to True, another widget will appear to choose beyond how many standard deviations a data point will be considered an outlier.
If True, will fit a smooth LOESS curve through scatter plot points.