Demand
KPN's ambition is to accelerate the rollout of the fiber optic network in the Netherlands by extensively automating the design process. However, due to overlapping or missing registration of route data, the quality of the historical data needed for this left quite a bit to be desired.
The process
Datacadabra has developed a cleanup of the telecom route using data science and machine learning, enabling further automation of design processes for construction and maintenance. To this end, we looked at route optimization based on cost minimization that included the unnecessary interruption of cables.
Data
The data that Datacadabra cleans up comes from KPN's databases. This data is merged with data from the Large-Scale Technology Registry (BGT), which also includes the subsurface (grass, asphalt) under which the cables were laid.
Services
As a solution, Datacadabra automates the pattern of traces to speed up KPN's design algorithms. The algorithm runs based on an application on KPN's system. This algorithm produces so-called gross gullies along roads, removing irrelevant gullies and adding new ones to be constructed.
Feedback
KPN can use the results of the algorithm to purposefully lay new fiber optic cables.
The result
The result is a basic design for a gully pattern that serves as input to KPN's design algorithm, which can automatically create a high-level design. This significantly reduces KPN's design time and accelerates design capacity. In this way, all of the Netherlands can enjoy the benefits of fiber sooner.
