Embed Size (px)
Citation preview
Fans just wanna have fun!
@petricek
Every year: hundreds of millions tickets sold worth Billions of Dollars
!
Recommendations!
!
Recommendations!
!
Who are my fans? !
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
Quality evaluation
Data collection
Recommendation Strategies
Quality evaluation
Data collection
Recommendation Strategies
1px
1px“pixel”
Javascript request URL:http://pixelserver:8080/pixel? evt=page_load imp=e4beb325-310e-493a-88cd-3da709ef509a impp=2d60a5ee-ac02-46df-8d0e-fbbe976361ef bid=bj6LvMJg-NnIOXp7tScUAHp-vcLsATCZoDvp08idDaIZo-4Iok1ppkDXZvA3RUWwoSV9IiRBqr05lhQRwpbH pid=lILEaTVwqW8iwByWQzh0Iuk3w4kxXAhBzoFx2-PBdBriuqOKeRousDZgIgrVZQ mid=lILEaTVwqW8iwByWQzh0Iuk3w4kxXAhBzoFx2-PBdBriuqOKeRousDZgIgrVZQ sid=lS9oSQ72XqeVxPX0axzEjwFMJQ5RfAbYOdA5G_IssvgmOl4fiQHQ5gO0Lg_FDQNZ5hirrvmiMwHlEcuh pgt=Artist%3A+On+Sale%3A+List dmn=TM_US aid=1687536 mkt=35 tsu=1433631937802 ref=http%3A%2F%2Fwww.ticketmaster.com%2F pgn=1 tml=tm_homeA_b_10001_1 fbt=not_authorized
1px
1px“pixel”
Javascript request URL:http://pixelserver:8080/pixel? evt=page_load imp=e4beb325-310e-493a-88cd-3da709ef509a impp=2d60a5ee-ac02-46df-8d0e-fbbe976361ef bid=bj6LvMJg-NnIOXp7tScUAHp-vcLsATCZoDvp08idDaIZo-4Iok1ppkDXZvA3RUWwoSV9IiRBqr05lhQRwpbH pid=lILEaTVwqW8iwByWQzh0Iuk3w4kxXAhBzoFx2-PBdBriuqOKeRousDZgIgrVZQ mid=lILEaTVwqW8iwByWQzh0Iuk3w4kxXAhBzoFx2-PBdBriuqOKeRousDZgIgrVZQ sid=lS9oSQ72XqeVxPX0axzEjwFMJQ5RfAbYOdA5G_IssvgmOl4fiQHQ5gO0Lg_FDQNZ5hirrvmiMwHlEcuh pgt=Artist%3A+On+Sale%3A+List dmn=TM_US aid=1687536 mkt=35 tsu=1433631937802 ref=http%3A%2F%2Fwww.ticketmaster.com%2F pgn=1 tml=tm_homeA_b_10001_1 fbt=not_authorized
Of Monsters and Men
1px
1px“pixel”
http://pixelserver/api/pixel?…
http://pixelserver/api/pixel?…
http://pixelserver/api/pixel?…
http://pixelserver/api/pixel?…
Number of requests over time
Number of requests over time
something big went on sale
Number of requests over time
something big went on sale
~10 x
Quality evaluation
Data collection
Recommendation Strategies
Quality evaluation
Data collection
Recommendation Strategies
Ticket Alert
~100m users
tens of thousands artists
~100m users
tens of thousands artists
~100m users
thousands of artists
~100m users
thousands of artists
Availability
Recommendability
Online Recommendations
Top Sellers
Offline aggregate over many days
Hot right now
30min sliding window
Hot right now
30min sliding window
count rank “hot right now”filter
After viewing
fans eventually bought
Personalized recommendations
100m users
tens of thousands artists
100m users
tens of thousands artists
100m users
tens of thousands artists
100m users
tens of thousands artists
100m users
tens of thousands artists
100m users
Recommender!Service
Service endpoints
Public endpoints
Ticket!Availability !
Service
Event Feed
Recommender!Service
Pixel!Service
100ms SLA
async
Storm
$ docker-compose up!:-)
$ docker-compose up!:-)
$ docker-compose up!:-)
$ docker-compose up!:-)
Quality
Expert feedback
Availability
A/B testing
Stealth testing
hdfs bolt
hdfs bolttee bolt
Recommender!service
replay!bolt
hdfs bolt
hdfs bolttee bolt
Recommender!service
replay!bolt
hdfs bolt
recA (live)
clickA
impA
recA (live)
clickA
impACTR
recA0
0.25
0.5
0.75
1
recB
recA (live)
clickA
impACTR
recA0
0.25
0.5
0.75
1
recB
recA (live)
clickA
impACTR
recA0
0.25
0.5
0.75
1
recB
recA (live)
clickA
impACTR
recA0
0.25
0.5
0.75
1
0
0.25
0.5
0.75
1
recB
Quality evaluation
Data collection
Recommendation Strategies
Contextual banditsNext steps
Contextual bandits
Exploration v Exploitation
Next steps
Contextual bandits
Exploration v Exploitation More context (user, artist, venue, event metadata)
Next steps
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
$$$
$$$
$
$
$$$
$$$
$
$
$$$
$$$
$
$
$$$
$$$
$
$
Counting unique users
Counting unique usersO(N*log(N))
Counting unique usersO(N*log(N))
O(N)
Counting unique usersO(N*log(N))
O(N)
HyperLogLog
Counting unique usersO(N*log(N))
O(N)
HyperLogLog
O(N)
HLL
$$$HLL
HLL
HLL
$$$HLL
HLL
HLL
$$$HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
map
HLL
HLL
$$$HLL
+
+
+
reduce
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
HLL
map
HLL
Unique female users ——————————
Unique users total
Unique users with family ——————————
Unique users total
Unique high income users ———————————
Unique users total
$$$HLL
=
HLL
HLL
HLL
HLL
=
=
HLL
SF Philharmonic
Phish
Meshuggagh Backstreet Boys
Chvrchs 15% Asian
Chvrchs 15% Asian K. Michelle 57% African American
Chvrchs 15% Asian K. Michelle 57% African American
Ramon Ayala 92% Hispanic
Chvrchs 15% Asian K. Michelle 57% African American
Ramon Ayala 92% Hispanic Tom Petty 96% Caucasian
Millennials Grits and Biscuits
Gen X Depeche Mode
Baby Boomers Chicago Musical
Tens of thousands models
Next
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
0
500
1000
Price
0
500
1000
Price
|+
|+
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vowpal wabbit
JNI
Storm
vw
vw
vw
vw
?
vw
vw
vw
vw
?
vw
vw
vw
vw
?
vw
vw
vw
vw
?
vw
vw
vw
vw
?
In-cart conversion
4 x
Arms Race
Next steps
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
!
Tickets to the Fans!(not scalpers)!
!
Recommendations!
!
Who are my fans? !
Now what?
Decades of granular and
longitudinal data
Now what?
Recommend the best Seat
for you
linkedin.com/in/petricek
