RhythmCount

Analysis of rhythmicity in count data has become an important aspect in different fields from science and engineering to economy and process planning. Several methods have recently been implemented to investigate the rhythmicity of continuous data. However, in most cases, these need to be manually adapted to work with count data as well. Namely, non-negative integer data that are usually obtained by counting of specific events. Herein, we describe the implementation of RhythmCount, an open-source Python module specifically devoted to rhythmicity analysis of count data. RhythmCount combines the cosinor regression model with different count data models. The proposed implementation allows automatic identification of the most suitable model for a given dataset, assessment of different measures and parameters of the rhythmicity of the dataset, and production of publication-ready figures that can be used for a straightforward interpretation of the obtained results. We demonstrate an application of the proposed module in the analysis and comparison of the daily traffic trends during the COVID-19 epidemic with the daily traffic trends in normal (non-epidemic) conditions. RhythmCount is available at https://github.com/ninavelikajne/RhythmCount under the MIT license. The implementation reported in this paper corresponds to the software release v1.1