This example shows changes in German voting behavior between 2015 and 2017 with a Sankey diagram, tracing how voters moved among parties and categories like non-voters and first-time voters. Party affiliations are color-coded on both sides of the flow, with tooltips revealing exact voter counts. The visualization uses the d3-sankey plugin to generate the diagram, alongside d3.v4 for general data handling and SVG rendering. Data is loaded from a local JSON file (poll.json) and processed to construct nodes and links based on party names and vote counts.