Random survey for crowd-sourcing similarities

Hi all
I have a data set of around 5,000 items that I need to test how similar or dissimilar people see each item as being from each other. In theory, the items should be able to be grouped up into major and minor groupings, providing the basis for a dendrogram structure.
To do this, I'm planning to develop a survey to crowd-source the answer. Essentially I want to ask individuals to rank a list of 10-20 random items about how similar/dissimilar they are, as well as some questions about why they grouped in this way. I am hoping that I can then use this to pull out the groupings. Appreciate that I need a lot of responses to do this, but I'm anticipating being able to fund some cash prizes to encourage participation.
  • Is anyone aware of a platform that could be used to run the survey? I usually use surveygizmo or surveymonkey, but these don't seem to allow me to generate a random subset of the 5,000 list to generate surveys.
  • Does anyone have any views about how the analysis might be done? I was planning to us Excel to pull this all together, and am interested if something else (R, Tableau etc) might be more suited to the job.
  • Has anyone done anything like this previously? If so, how did you go about it, and how would you suggest approaching this?