Analyzing Late Antiquity Shifts of Trade Regime in the Iberian Peninsula and Their Causes via Change Point Detection Methods

History attempts to make sense of disparate information trying to create a discourse that lays a series of events with crisp cause-effect relationships in a sequence. Epochal shifts, such as the change from Antiquity to Middle Ages, are especially complex since they involve a large number of economic, political and even religious factors which occur over long periods and that might overlap and interact through reciprocal feedback mechanisms, making this cause-effects sequence difficult to establish. In this research we adopt a data-driven and well-established methodology to identify, with quantifiable statistical precision, the moment when this shift happened, and from there arrive at its possible causes. We will use historical coin hoard data to find out whether such a shift is detected in a peripheral part of the Roman Empire, the Iberian Peninsula. To do so, we will apply different changepoint analysis methods to a time series of trade links created from that data, and conduct a retrospective analysis based on that result, analyzing the structure of the trade networks before and after the link. Thus, we progress from identifying when the shift happened to identifying where it took place, which in turn allows us to get to investigate why it happened; namely, historical events that could have caused it. This methodology can be used to analyze epochal changes in several steps using time-stamped network data, possibly finding disregarded causes or cause-effect links that could have been overlooked by qualitative methods; in this case, we have applied it to a dataset of coin hoards either found in the Iberian Peninsula or including coins minted there, finding a changepoint in the early 5th century, which, through network analysis, has been linked to a loss of trade with the area of Britain.