A large body of Australian laboratory test data is or will be available for the development of a new high resolution traffic emissions prediction tool. Consideration of data quality is an essential step in the development of this empirical model. Several potential issues are discussed in this paper. Although the majority of issues can be dealt with either before, during or after model development, not all issues can be addressed due to a lack of information or empirical data. This is not a problem specific for Australia: international research is ongoing to address these issues. As part of an ongoing process of model improvement, the relevance of gaps in knowledge need to be further explored and eventually addressed.