Nifty New Netflix Tracker Report


I don't know if this will push any of you over the hump into using the Netflix Tracker, but I just launched the Turnaround Time reports. Here's a sample of what your turnaround log might look like:

Movie ReturnedNext Movie ReceivedTurn-
aroundInherit the Wind
returned to SA on 8/16/2004Jesus' Son
received on 8/19/2004 from SJ 3Diamond Men
returned to SA on 8/2/2004Inherit the Wind
received on 8/6/2004 from SJ 4Auto Focus
returned to SA on 7/23/2004The Deep End
received on 7/27/2004 from SJ 3The Good Thief
returned to SA on 8/13/2004The Believer
received on 8/19/2004 from SJ 5AVERAGE:3.75
I'm particularly pleased with this report (and the one that aggregates by center, linked to above) because it was a huge pain to program the progression of what movie ships out in response to a given return.

I appreciate all your hard work on this tracker. I'm impressed that you were able to design this turnaround report, and I look forward to getting my results.

Thank you! I'm really happy this one seems to be working, as it was pretty tricky. There is well-known code in SQL (database set-oriented language) for computing time gaps, but it only seems to work on nice sequential data. Your Netflix history has movies returned and received on the same day. So you might send movies back on two successive days, but Netflix might get them on the same day. Or Netflix might get multiple movies on the same day, but might send replacements over the next two days. I couldn't find a way to do it in straight SQL, so ended up looping over the complete history, basically trying to fulfill this rule:

"To figure out which movie is sent as a replacement for a returned movie, use the first movie shipped on or after the day the returned movie is received, unless that movie is already associated with another return, in which case find the next movie shipped..."

...and so on and so on, until you've matched up everything you can. Unfortunately this process is too involved to run each time somebody views the report, so I rebuild that "movie progression" nightly.

Anyway, I bet that's more than you wanted to know. Glad you dig it!

No, I think it's fascinating. I use SQL here at work, but not to the same extent you do. I use it in an Access database. I know that you're using a database to serve us our Web pages here on Listology, but I have only a rudimentary knowledge of how that might work. I have some very basic programming knowledge, so I appreciate your efforts.

You finally convinced me. I've joined the tracking list, though I probably won't join the queue lists (since I think it's weird if people knew what movies I was renting).

:-) Excellent!