The critical node detection problem is a challenging computational task in computer science, with applications in biology, network risk analysis, social network analysis, etc. Generally, the problem consists in finding a set of k nodes that, if deleted, will maximally degrade the network according to a given measure. The main goal of the paper is twofold: to introduce a new measure which can be used for critical node detector in weighted undirected networks, and to use this new introduced measure in a data analysis task. Numerical experiments are conducted on synthetic networks, and as an application, the problem is used for data analysis.