Clustering is an important unsupervised learning algorithm that groups records according to their similarity. However, since uncertainty has become an inherent of real world datasets, crisp clustering leads to inefficiency. Hence, introduction of uncertainty based models like the fuzzy set and the intuitionistic fuzzy set is necessary to compensate for the ambiguity in data. Huang, in his fuzzy k-modes algorithm, introduced the fuzzy component in clustering categorical data by modifying the existing k-means algorithm. This correspondence describes an intuitionistic fuzzy k-modes algorithm for clustering categorical data and establishes it to be more efficient than the fuzzy k-modes algorithm. Metrics like accuracy, DB index and Dunn index are used to compare the efficiency of the two algorithms. The experimental analysis section shows that the proposed algorithm is more efficient than the existing one. Several graphical and tabular representations have been provided for easy comparison of the results. © Springer Nature Singapore Pte Ltd. 2017.