DJ techniques are an important part of popular music culture but are also not sufficiently investigated by researchers due to the lack of annotated datasets of DJ mixes. This paper aims at filling this gap by introducing a publicly available DJ-mix dataset. It contains automatically generated beat-synchronous mixes based on freely available music tracks, and the ground truth about the placement of tracks in a mix. Each mix is generated in several variants with different effects and time-stretching methods applied. Possible applications are to test novel methods for track identification in mixes, or for automatic annotation and deconstruction of recorded mixes for which the constituent tracks are known.