Companies are increasingly embedded in B2B environments, where they have to collaborate in order to achieve their goals. Such collaborations lead to inter-organizational business processes that may be commonly supported through the exchange of electronic data interchange (EDI) messages (e.g., electronic purchase orders, invoices etc.). Despite the appearance of XML, traditional approaches to EDI, such as EDIFACT and ANSI X.12, still play an overwhelmingly dominant role. However, such traditional EDI standards lack a notion of process. In other words, the exchanged business documents are typically not embedded in the context of other exchanged business documents. This has two shortcomings: (1) the inability to apply proven business process management (BPM) methods, including process mining techniques, in such settings; and (2) the unavailability of systematic approaches to business intelligence (BI) using information from exchanged EDI messages. In this article, we present the EDImine Framework for enabling (1) the application of process mining techniques in the field of EDI-supported inter-organizational business processes, and (2) for supporting inter-organizational performance evaluation using business information from EDI messages, event logs and process models. As an enabling technology, we present a method for the semantic preprocessing of EDIFACT messages to exploit this potentially rich source of information by applying state of the art BPM and BI techniques. We show the applicability of our approach by means of a case study based on real-world EDI data of a German consumer goods manufacturing company.