You could do this a couple of ways. If you have enough sequence data you could try to assemble the genome and then find the contig with the kanamycin cassette with a blat/blast search. Or you could try to bin the reads containing kanamycin cassette using bbsplit.sh from BBMap suite. Then try and assemble those reads using a k-mer based aligner such as tadpole.sh again from BBMap suite (guide here). Then look at the reads (past the edge of kanamycin cassette) to see if you can identify where the insertion is located in the genome.