In book publishing, people are always trying to forecast trends so that they can get in front of them. They are trying to answer the question "What's the next big thing?" For this proposal, I've been working on researching PAST trends -- and in hindsight there are definite cycles that trends go through.
yes
So, if you shift your goal from trying to predict a trend to identifying a trend -- you have success (especially since there is a sweet spot where if you can hop on a trend fast enough, it produces great - if not amazing - results) So - I want to create a model to forecast the "curve" of trends to identify the best ones and the sweet spots...
The majority of books on amazon are tagged and "trends" would need to be searchable by tagged subjects.
I'd like it to follow:
1. Groups of novels (fiction only) tagged with similar subjects.
2. Sales figures for those novels over the lifetime of the book (Amazon rankings are fine, or Bookscan numbers)
3. Be able to chart when a trend has: a) hit it's high point, b) it's point of saturation - before new books with the same tags no longer sell as well
and I'd like it to be able to regularly pull information off the web (Amazon, most likely) to automatically update the material.
--
Other information:
Rights you are purchasing (full copyright, internet rights, etc.): Full copyright.
Payment terms (up-front, milestones, etc.): milestones (half on execution of 'contract,' half of delivery and acceptance)
Payment methods (PayPal, escrow, etc.): PayPal works, or we can make alternative arrangements.
Again, thank you in advance - I apologize for any lack in this description. I look forward to seeing your bids.
Nadia
I am a domain expert in data mining and content extraction. I have worked on extracting data out of Amazon for several customers. I already have the code base for crawling Amazon for product reviews and data. I can send you a sample set of the data captured.