Michael-Diamond-Data
commented
7 years ago Thanks for the comments. I like the idea of automating the customer input a bit more. Might be some simple tradeoff type survey - where people identify first some movies they've seen, then rate which ones they liked (A vs. B) with perhaps some criteria around why they liked them.
Will edit for clarity and flow. Was perhaps a bit stream of consciousness, hence the parentheticals!
Hey Michael, Nice job identifying a real, universal problem! TV show/movie selection is so difficult, and it doesn't make sense for it to be. There is much data available, as you reference, so automating the decision seems like a great idea.
I thought your write-up was clear, if a bit long. The examples you provided were very helpful to define the terms you're using. I did find the parentheticals a bit distracting at times, though. Perhaps you could discuss a single example in great depth early on, and then reference back to it from the rest of the proposal? Rather than having to give examples in each sentence.
For data collection: the end of the Key Units & Metrics section says that "individuals would need to provide information on their preferences, previous viewing history, actors and directors, subject matter, etc". It sounds funny, but I'm not sure I actually could identify those things as a viewer. I know which shows & movies I enjoyed, but wouldn't recognize whether there is a common director between them. Is there a way to automate that pattern recognition as part of the data collection process/interface? Maybe do a sort of tinder for video content --- a viewer just marks yes/no for whether they liked it, and rates many shows very quickly. Then the program can find similarities.
Overall, this is a solid proposal. The problem is easy for readers to relate to. You identified your customer as the video provider in Section 1, and really stuck with that customer view throughout the document. (Impressive focus!) Very interesting idea!