In my opinion, the first step to start is reading Steinar Kvale – his books and papers. It is mostly about unstructired interviews but very helpful in your field too. Good luck!

thanks Emilia will have a look

I pubblished a papper few years ago proposing a method for that "methodological proposal for analise qualitatite data in two levels.

Johnny Saldana's methods for analyzing data is quite helpful. Norman Denzin's work is quite helpful as well.

If you are intending to do Grounded Theory you would code and analyze structured and semi structured data alike. No challenges of the form here.

As Jordan said, Norman Denzin provides a good source for methodology.

If what you have are semi-structured interviews, then it probably means you have an hypothesis you're testing. I could be wrong, you could be using, for example, a modified YLS/CMI to construe new theory.

In either case, look at each interview in full, and start coding, that is, assign a subject to each section of text. If you're working on an hypothesis, then you already have your subjects to assign. Otherwise, you'll need some interpretation (but please be consistent) while reading those transcripts, and then assign a meaning/theme/topic to each segment. Since these are semi-structured interviews, you will run across the same topics at each interview and, hopefully, get different perspectives on each topic.

Now, how do you report qualitative data on the final work?

This, again, depends on whether you're working on testing an hypothesis or on grounded theory. In the first case, you'll use excerpts from the transcripts to strengthen and verify said hypothesis; on the second case, look into the topics you coded before, and try to find the subjects' perspectives on each of them. Looking beyond that, if the subjects belong to a specific community or cultural group, their answers are a cue to their community's or group's ethos and beliefs. This, of course, will require some interpretation, but is that bread and butter of qualitative research!

Best of luck and good work! =)

For analyzing semi structured interviews, thematic analysis should be done.