Process mining relies on activity logs to discover process models, check their conformance, enhance processes, and recommend the next activity. On another side, many environmental factors such as time, location, weather, and profile are obtained from many sources, such as sensors, external systems, outside actors, or domain knowledge bases, and could also enhance recommendations. The existing research mainly focuses on single activity log datasets; only a few consider combining various sources. Our main goal is to provide better inputs to process discovery and better recommendations. In this paper, we focus on the combination of activity logs and sensors data with domain ontology as an intermediate step to attaining our goal. We use a case study of smart home activities to test this combination.