Truncating data to speed up process flows or reduce storage space is quite common in organisations. In many applications, the decimal precision of numerical data is considered of limited importance, particularly when precision extends beyond the commonly used two decimal places. However, this assumption can introduce significant challenges when modelling complex, high-volume production data. This thesis is motivated by practical difficulties encountered in modelling Electronic Data Interchange (EDI) production data within a supply chain organisation. EDI data is bursty in nature, presenting challenges for modelling, therefore this work develops a structured framework for modelling EDI data. It investigates different approaches to identifying quantisation, ways to overcome quantisation challenges, and what impact, if any, it has on modelling applications. Finally, the research explores techniques for recovering information lost through rounding, addressing the problem from both a data point reconstruction perspective and a distributional recovery perspective.