Recommender systems experience the challenging and complex process of finding educational resources recommendations for teachers which requires a careful awareness of the teacher's context. These systems should take into consideration the contextual coexistence of teachers in multiple environments or contexts such as social, working, and sentimental contexts. The collected data for each of these contexts is assembled from different sources and systems with different data architectures. Accordingly, this paper provides a multi-step proposition towards a personalized educational resource recommendations for teachers. Furthermore, the paper discusses the data sources from which the contextual data is collected, in addition to the purposed methodology for integrating and mapping the data into the ontology. At the end, the paper illustrates the importance and the role of the teacher analytics and the sentimental data in this proposition. The proposed approach results in a complete representation of the teacher's context and hence, the personalized recommendations.