The data collected by a gravitational wave interferometer are inevitably affected by instrumental artefacts and environmental disturbances. In particular, for continuous gravitational wave (CW) studies it is important to detect narrow-band disturbances (the so-called "noise lines") during science runs, and to help scientists to identify and possibly remove or mitigate their sources. The NoEMi (Noise Frequency Event Miner) framework exploits some of the algorithms implemented for the CW search to identify, on a daily basis, the frequency lines observed in the Virgo science data and in a subset of the environmental sensors, looking for lines that match in frequency. A line tracker algorithm reconstructs the lines over time, and stores them in a database, which is made accesible via a web interface. We describe the workflow of NoEMi, providing examples of its use for the investigation of noise lines in past Virgo runs (VSR2, VSR3) and in the most recent run (VSR4).