Once you find a song, grab a piece of paper and make a graph, with energy levels on the X axis and time on the Y axis.
As the song progresses on the Y axis in time, place a dot on the energy level you think the song projects.
Afterward connect the dots to form a line graph.
Do this a few times until you're satisfied with the graph. Now you'll have a good representation of the energy levels of your music, so now you can grab your video footage to best match your music. Most cases will have fast moving video to accompany high energy levels of music, while low energy sound will be played with slow moving footage. As the sound energy builds up, you can use that with video footage that shows a buildup of a fight or another type of event.
Hope this makes sense...and as always, do what you think is fun.