First you choose a subject for the analysis. Then you review all possible literature on the subject you chose. Then, you establish the aspect you are interested in investigating. Then you should prepare a coding sheet for data collection. The coding sheet should include the variables you are investigating. You are collecting the data systematically into the coding sheet, and in the final stage you are analyzing the data using some non-parametric statistics.
Recommended: while choosing the variables for the investigation, you also establish the study's research question.
1) Organize your data (transcribe it, check it for accuracy, get familiar with it)
2) identify your framework (are you using an existing model or theory, or creating your own?) Your framework is your coding plan. Are you using inductive or deductive coding? Are you going to be guided by your questions, or by the data you get?
3) Sort your data into the framework. Code the data, adapting your framework as you go if necessary. (Eg adding questions/themes that arise from the data that weren't in your existing coding scheme)
4) Use the framework to shape your descriptive analysis, identify recurring themes etc.
5) Conduct a secondary analysis where you look at themes, how they interact and link together, and impact each other.
I am not sure that it is necessary to do any non-parametric statistical analysis if you are doing a qualitative study.
Why do want "five" specific stages -- is there a reference that you are trying to find? In may experience, there are many different versions of qualitative content analysis, that could easily be broken into four or five or six or whatever stages.
Please refer to Schilling’s (2006, pp. 29-35) qualitative content analysis spiral incorporating a 5-level of analytic process.
Schilling, J. (2006) On the Pragmatics of Qualitative Assessment: Designing the process for Content Analysis, European Journal of Psychological Assessment, 22, 1, pp. 28-37.
Almost all of the versions described above allow you to begin with a set of prior codes that you can use deductively, and then add new codes inductively. This is known as a hybrid approach.
There is not one framework to use. You can create your own. The conceptual framework you use is your own approach to exploring the research problem or question. you group your findings and themes into an overarching framework that encompasses your data and findings. I am not sure what you mean by framework since there are many definitions and examples. Hsieh and Shannon have a great article outlining three approaches to content analysis depending on your approach. maybe this will help? https://www.ncbi.nlm.nih.gov/pubmed/16204405