This example visualizes a force-directed network where nodes represent social media users and websites, with node size mapped to the volume of shared links. Built with D3 v4, it uses a custom collision function to prevent overlapping circles while the simulation runs. Each node’s radius reflects the total `link_share` value aggregated from the dataset, and the force layout spaces nodes while respecting collision boundaries. The visualization is rendered in SVG and shows how different users cluster around the news sources they share, with node size making relative activity visually comparable at a glance. The collision behavior adds clarity by preventing overlaps, making the network easier to explore. Data comes from a CSV listing website, user name/ID, and share counts. This example is a fork of jinniluo’s block, adapted with the D3 v4 API.