Create a Python script to analyze posting data stored in JSON format

I need help in analyzing data from around 90 online forums as part of a research project - I want to identify factors that predict the success of online forums. Access to an Amazon AWS instance will be supplied as the files are over 30Gb. Work will need to be done via the Amazon AWS instance. Ideally we will be able to use the Python code in an iPython notebook. I am flexible on this however. The data comes nicely packaged in a set of JSON files: [1] [login to view URL] Contains badge data associated with user accounts - i.e. the user ID it was assigned to, the name of the badge, and the date/time it was assigned. [2] [login to view URL] Contains comments data - e.g. timestamp, user ID of the user writing the comment, score (upvotes/downvotes). [3] [login to view URL] Contains posting history data - e.g. posting revision type, details of the changes made, user ID associated with the user who made the change [4] [login to view URL] Contains posting data - e.g. the actual questions and answer text posted by users, tags associated with the questions/answers [5] [login to view URL] Contains user data - e.g. user ID, user account creation date, last access date, reputation score, etc. [6] [login to view URL] Contains voting data - e.g. vote type ID, timestamp of the voting action, user ID of the person making the vote A data dictionary with more details is attached as is a sample copy of the JSON files (only containing 3 forums worth of data to make it easy to download - the real version has around 100 forums). In brief - [1] we will need to classify each of the forums into two groups: success and failure. I will tell you which ones failed and which ones succeeded. [2] We then focus on the first 14 days of data for each forum. [3] We then calculate metrics for each one - e.g. average number of posts per day, average number of answers per day, concentration of people doing all the work [4] Make conclusions based on the results using a linear regression or other model