This page is from our blog archive
It's still useful, but it's worth searching for up-to-date information in one of our more recent blog posts.
- 32 million business and political relationships analyzed and visualized using KeyLines and Neo4j
- First KeyLines graphs produced after just two hours with the toolkit
- A fully functioning KeyLines GUI built and released in one month
IntroductionIn business, understanding connections between people and institutions can be the key to success. A clear picture of these networks of influence can provide marketers with connections to thought leaders, sales teams with a direct route to decision-makers, compliance managers with more effective due diligence and researchers with a wealth of previously buried intelligence. Kantwert GmbH, founded in April 2014, is leading the way in making this data available and navigable to whole new audiences. Their unique data analysis and visualization platform allows users to search and traverse networks of directors, politicians and institutions in Germany. The database uses rule-based logic to detail 32 million relationships between more than 3.3 million people. This case study explains how Kantwert has used the KeyLines toolkit to bring even greater clarity to these networks, providing an intuitive way to interpret and navigate the complex connected data.
The ProblemThe Kantwert team approached Cambridge Intelligence in January 2015. They already had a product with pilot customers, but were looking to upgrade the visualization component sitting at the heart of their GUI. The existing visualization, which was built in-house, suffered from bugs and sluggish performance. In particular, the pilot customers had commented on the inefficient automated layout and inability to move nodes around the chart. This limited their ability to find insight in the data. At the back-end, the platform uses a Neo4j graph database to store and traverse the connections, plus a Cassandra database to store the content for each data point, so compatibility with these systems was essential.
The RequirementsThe Kantwert team had identified a number of criteria for their visualization solution:
- The ability to visualize complex connected data in a web browser
- Compatibility with all modern browsers, including legacy browsers back to IE8
- Full native support for the Neo4j graph database and Cypher queries
- Full visual customization – e.g. chart branding, icon styles, node glyphs, etc
- An effective force-directed automated layout
- Draggable nodes and links
- A customizable event model, to define behavior around interactions (e.g. double-click, right-click, etc).