md`Trace the dataflow graph to see how calculations are carried out. For example, given the input dataset \`table\`, an \`extent\` operator calculates the minimum and maximum values of the \`value\` field to determine the \`yscale\` \`domain\` parameter. Similarly, a chain consisting of \`aggregate\`, \`collect\`, and \`values\` operators calculates unique \`key\` field values to determine the \`xscale\` \`domain\`. The \`xscale\` operator is also parameterized by the \`scalePadding\` signal. The \`table\` dataset is encoded as rectangle marks by a sequence of operators: - A \`datajoin\` operator maps input data objects to scenegraph mark items. - A \`collect\` operator gathers the output of the data join. - A \`mark\` operator binds the generated mark items to the scenegraph. - An \`encode\` operator computes visual encodings to determine mark item properties, and is parameterized by \`xscale\` and \`yscale\`. - A \`bound\` operator computes bounding boxes for each mark item. The bounds are used for both layout and input event hit testing. - A \`render\` operator queues scenegraph updates for (re-)rendering. - Finally, a \`sieve\` operator propagates mark updates (but not the mark items themselves) to other parts of the dataflow graph, causing downstream operators to be re-evaluated if and when any mark items change. In this case, the top-level \`viewlayout\` is re-run when mark items change. The other set of visual encoding operators — starting from the \`root\` node — calculate properties and layout for the visualization's top-level group mark. `