Abstract: In one example, a processor may generate a graph of vertices representing distribution areas of a telecommunication network and links between vertices of adjacent distribution areas, calculate similarity measures for the links based on distance scores between vertices joined by the links relating to: percentages of aerial infrastructure and percentages of multi-unit buildings in distribution areas represented by the two vertices, and whether or not the distribution areas represented by the two vertices are on a same route from a central office of the telecommunication network. The processor may label the links with the similarity measures to create a link-weighted adjacency graph, perform community detection on the link-weighted adjacency graph to generate sub-graphs, each sub-graph including at least one of the vertices, and provide a map of construction zones based upon geographic areas covered by distribution areas associated with vertices in the sub-graphs.

Abstract: In one example, a processor may generate a graph of vertices representing distribution areas of a telecommunication network and links between vertices of adjacent distribution areas, calculate similarity measures for the links based on distance scores between vertices joined by the links relating to: percentages of aerial infrastructure and percentages of multi-unit buildings in distribution areas represented by the two vertices, and whether or not the distribution areas represented by the two vertices are on a same route from a central office of the telecommunication network. The processor may label the links with the similarity measures to create a link-weighted adjacency graph, perform community detection on the link-weighted adjacency graph to generate sub-graphs, each sub-graph including at least one of the vertices, and provide a map of construction zones based upon geographic areas covered by distribution areas associated with vertices in the sub-graphs.