I’ve had a hard time wrapping my head around knowledge graphs. As I explained in this earlier post, graphs themselves are not new, and the overall structure of a graph is somewhat intuitive and easy to understand. But how do we get from sticks and balls to providing valuable business insights and performing complex data operations with promises of staggering value and efficiency? 

Let’s get one step closer to bridging the gap by taking a look at three knowledge graphs you already interact with on a regular basis. These knowledge graphs are virtual representations of real-life systems. You might be surprised to discover that you know more about knowledge graphs and how to use them than you think!

Power Grids

Power grids are complex networks that fit the graph model of nodes and relationships. Power producers, consumers, and transformers can be represented as nodes, and the connections between them are relationships. 

Source: Wikipedia

A complete knowledge graph of a power grid can be used as a central information resource for the power producer, line operators, service personnel, and customers. Questions such as “which houses will experience an outage if this transformer is taken offline?” are easily answered by a graph query. 

In China, researchers demonstrated that implementing a knowledge graph and artificial intelligence could improve the reliability and efficiency of China’s large and aging grid. The abstract states that “through a series of experimental simulations, we show that the efficiency of daily operations, maintenance, and management of the power grid can be largely improved.”

Source: Knowledge Graph Construction and Application of Power Grid Equipment

Public Transportation Routes

Any public transportation system is a graph. Stations are represented by nodes, and the routes between them by relationships. Station nodes might have attributes such as name, location, and wheelchair accessibilities. Route relationships might have attributes such as frequency and speed. You are performing an operation on a graph whenever you look at a transportation map to figure out the shortest route between two stations.

Public Transportation Routes

A public transport knowledge graph could be a central resource for all departments involved in creating, running, and maintaining transport lines. This knowledge graph can also be mined using artificial intelligence and algorithms to solve problems and identify issues in the system. In Brazil, researchers created a graph-based data processing methodology that identifies discrepancies in supply and demand for the local bus network, as well as bottlenecks.

Source: Graph mining for the detection of overcrowding and waste of resources in public transport

Cell Networks

All digital networks are essentially graphs. Each one of our cellular devices is a node, connected to cell towers via a wireless relationship. Our devices and the cell towers are nodes that have attributes, such as location, device type, antenna capability, or carrier. These nodes are connected by a wireless relationship that can also have attributes such as strength and connection type (e.g. 1X, 3G, 5G). The cell towers are also connected to other networking devices through various connection types. Here is a graph schema that shows the potential types of nodes (devices) and relationships (connections) in a 5G cellular network.

Source: A Survey of 5G Network: Architecture and Emerging Technologies

Similar to the examples above, a mobile network knowledge graph could be a central data resource for mobile network operators, tower maintenance personnel, and mobile network users. Researchers at Nokia and Huawei also demonstrated the effective analysis of a knowledge graph to provide holistic insights for mobile network operators. The authors state in the abstract, “In this paper, we propose graph theory based network insight analysis framework which can give mobile operators insight about their networks, provide better network coverage, and optimize [their] infrastructure investments.”

Source: Graph theory based mobile network insight analysis framework

What Does It All Mean?

What do power, transit, and cell service organizations have in common? Each of them can use their knowledge graph as a central source of important information. They can also leverage processing and analysis of the graph to gain crucial insights, for example:

  • Where is another power substation, cell tower, or subway stop needed?
  • How can we accommodate peak usage?
  • What is the impact of an outage or scheduled maintenance?
  • What is the best way to ensure service interruptions are minimal?
  • Are there redundancies? If so, are they necessary?

 

Max Swisher