Miscellaneous notes about other things that are
vaguely similar (or relevant) to UR8IT

UR8IT Main Page

Developers – Jump down to forms examples and discussion
USERS (beta testers) – Register and/or Login
Contact Scott Covert – 705 749 2225 – scott@rtopartners.ca

What about Amazon’s recommendation system?

I played around with it today (Aug 22 2015) and it’s got huge differences from the way it would be if it were integrated with a system like UR8IT. There’s a dozen ways to see individual products or lists on Amazon, and if you’re lucky enough to see “Why Recommended?” under an item, you can click on that and start adding scores. It’s not so easy to score products that have never been in your cart (or items related to those items), though. You can click “Improve Your Recommendations” to score every item that’s ever been in your cart (which is a circular and confirmation-biasy way to do things). Then you can click “Recommended For You” and further refine things. But from what I can tell, you can’t tell Amazon your opinions about stuff you’ve never bought at Amazon (or tell it that you probably would have hated some of the stuff in your cart that you didn’t buy), and then have it recommend a bunch of stuff from authors you’ve never heard of. It mostly just kept recommending more products from the same authors and bands I’m already a fan of. UR8IT is more about finding bulls-eye recommendations the user would not have easily found without the system.

Explicitly asking users their opinions for their lifetime favourites, and ignoring their carts, would, I believe, serve them better.

There’s no actual scoring system for predictions, and no in-line customized prediction of whether a user will enjoy a single item that they are viewing. Hence, completely different from the UR8IT vision.

JINNI.COM

See notes on the UR8IT home page.

MovieLens

A huge project done by the University of Minnesota – completely non-commercial. This claims to be based on collaborative filtering, in other words, “Taste Matching”, although of course they don’t disclose all the internal math. I assume their process involves putting users into taste groups then recommending movies based on users with similar tastes. As you know the UR8IT system works differently, not by matching users to users (at least not until there are 10s of thousands of users, and even then, only as 1 simple factor), but by matching users with their measured relationships (correlation strength and slope) to specific factors within a given medium and genre.

Criticker.com started in 2004.

Their goal is to use a single method to predict your scores for movies: peer match – based on other user’s average difference in scores on movies.

The site is visually noisy in an attempt to monetize through:
– paying $5 a month for ads-free site
– paying $1.50 one time to generate a thorough predicted score list through entire database
– ebay, etc.

The visual noise is a little more intense than UR8IT will be. Distractions are a popularity killer; they give the impression that the site is just a site, not a multi-platform service intended for tens of millions of people joined in the cloud.

The recommendation lists are hyperlinks you need to click to go to a separate movie page, in order to read more about it and submit a score. The UR8IT approach will be to provide simple lists for any requested medium and genre of requested recommendations, with boxes beside them for scoring, plus of course the predicted score if applicable.

TRAKT

I always panic when I see a new (or old) site or App that appears to be predicting entertainment scores the same way I envision.

But like the others, there turned out to be significant differences.

To start, Trakt is designed to be plugged into media box apps that are very often used for downloading movies and TV illegally.

Furthermore, once you join your Trakt account to a media box add-on like Kodi, you have to be careful about your settings or it will broadcast your entire current directory of (pirated?) movies and TV to the world at large.

To understate the case, this is not good.

The interface is far more visually and textually noisy than UR8IT will be, supported by ads (you can pay for ad-free interface) and other characteristics I plan to (or desperately hope to) avoid on UR8IT.

UR8IT will be created to interact with large, legal enterprises.

Contact Scott Covert – 705 749 2225 – scott@rtopartners.ca

Programming has already begun. If you are a programmer or business leader and want to be involved, please let us know asap. This section will very soon have references to our beta testing page and other resources.

After a test system is fully implemented, we will seek our first round of funding or crowd-funding and hire a professional statistician to improve the prediction formula, as well as hiring a database input team, and starting to market and advertise and meet with industry partners.

If we work together well you can be involved in the expansion. But first we need to build the simple basic system.

This is the form for UR8IT staff or contractors to add new movies, TV shows and more to a master database.

 



We are not showing this now, because we will be initiating the system based off of a spreadsheet of movies, not by adding them manually.

Contact Scott Covert – 705 749 2225 – scott@rtopartners.ca

We were able to integrate a lot of “future” components into v0.1, so we re-did the graphic below for a more optimistic development picture.

Components Of The Simplified v0.1 Of The Application

UR8IT system components