Catch and fishing effort time series are used by managers to safeguard the availability of resources in the future. Fisheries organizations estimate the status of a stock and the levels for sustainable harvest. Based on these indicators, regulations are developed to guarantee the availability of food and sustain economic growth. For example, tuna stocks are important not only in terms of nutrition, but also for the welfare, culture, revenue, and employment of the countries that rely upon them. However, fisheries time-series are the result of many natural and non-natural factors and are usually difficult to predict. Here, we propose a method to aggregate fisheries time series in space and time, to detect hidden periodicities, to predict the time series in the future and to identify the most stressed locations in the fishing area. We apply our method to tuna fisheries data in the Indian Ocean and compare it with respect to other methods. We demonstrate that the method is able to highlight periodic patterns in this high fishing activity area and to forecast the fisheries time series in the future. We use a research e-Infrastructure to comply with modern approaches supporting reproducibility, repeatability, and results sharing.