This work gives an insight if clustering technique applied to the dataset consisting of about 1000 measurements of High-Temperature Superconductors (HTS) using the AC susceptibility method, will allow recovering known and unknown relationships between different types of HTS and their superconducting properties, which depend on the type of superconductor and its preparation procedure. The dataset was simplified by using a Convolutional Autoencoder and the Bag of Words (BOW) representation. K-means and DBSCAN (Density-based spatial clustering of applications with noise) algorithms were used for clustering. The obtained results were visualised by the t-SNE algorithm (t-Distributed Stochastic Neighbor Embedding).