This demo shows how to use KeyLines to find the origin of fraudulent credit card transactions. The data is taken from a realistic Credit Card Fraud Detection scenario created by Neo4j GraphGist:

• Nodes represent people and merchants, with glyphs on merchant nodes to show transaction values.
• Red links represent disputed transactions; green ones represent undisputed ones.

There's two-way interaction between the chart and time bar to help you understand the data:

• The histogram shows total transactions, with disputed claim activity shown in red using the time bar's mark function.
• Selecting a node or link displays trend lines for that item - green for undisputed activity and red for disputed activity.
• Tooltips on trend lines display the number of transactions and their total value, while KeyLines uses ping to highlight relevant nodes so they’re easy to spot on the chart.
• Zoom into the time bar to focus on transactions that took place during the exact time period you’re interested in.
• Use the extend play control to show how the data evolved.

We’ve added a Select people with disputes filtering option, so it’s easy to focus on nodes linked to disputed activity.

Key functions used:

• chart.foreground
• chart.layout
• chart.ping
• chart.graph().neighbours
• chart.graph().degrees
• timebar.getIds
• timebar.inRange
• timebar.options().groups
• timebar.selection
• Chart Events
  • 'selection-change'
• Time Bar Events
  • 'change'
  • 'hover'