SlideShare a Scribd company logo
1 of 1
Download to read offline
SciPy and Real-time Big Data for Site Optimization
Pyleus
Message
Processor
Bolt
Pyleus
Event
Worker
Bolt
Pyleus
SciPy
Optimizer
Bolt
Pyleus
Update
Messenger
Bolt
SciPy
Bayesian
Bandit
Spout
Spout
Application
State
Visitors to Bankrate.com
Impressions
and Clicks
Improve User Experience
For more info, contact:
Winnie.Cheng@bankrate.com
Which
Variation to
show
Bankrate.com Data Science and Engineering Team
Example: Pick better story headlines
Objective for Site Optimization:
Enable fast and cost-efficient ways of testing
new designs to improve user experience
Algorithmically decide which of two headlines to
show user to maximize click-thru-rate (CTR)
Computation Framework with Kafka-Storm
Simulation Results
With more data, algorithm becomes more
confident of estimated CTR for each variation
Bayesian Multi-Armed Bandit
algorithm on Storm Topology
decides how often to show
each variation by analyzing
impressions and clicks
Iteration: 100
W1: 56.37%
W2: 42.63%
Iteration: 1000
W1: 9.82%
W2: 90.18%
Iteration: 2000
behavior reversal
W1: 64.06%
W2: 39.94%
Iteration: 3000
W1: 94.84%
W2: 5.16%
Iteration: 4000
W1: 97.12%
W2: 2.88%

More Related Content

Similar to Scipy_v1_reprint

Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scaleQcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scaleKinshuk Mishra
 
Vpon - 廣告效果導向為基礎的行動廣告系統
Vpon - 廣告效果導向為基礎的行動廣告系統Vpon - 廣告效果導向為基礎的行動廣告系統
Vpon - 廣告效果導向為基礎的行動廣告系統Vpon
 
AWS Activate Webinar - Growing on AWS
AWS Activate Webinar - Growing on AWSAWS Activate Webinar - Growing on AWS
AWS Activate Webinar - Growing on AWSAmazon Web Services
 
20141209 meetup hassan
20141209 meetup hassan20141209 meetup hassan
20141209 meetup hassanNanda Kishore
 
Leverage Real-Time Purchase Intent to Boost Sales & Customer Growth
Leverage Real-Time Purchase Intent to Boost Sales & Customer GrowthLeverage Real-Time Purchase Intent to Boost Sales & Customer Growth
Leverage Real-Time Purchase Intent to Boost Sales & Customer GrowthTinuiti
 
Criteo | Retargeting | BTO 2016 | Daniele Beccari
Criteo | Retargeting | BTO 2016 | Daniele BeccariCriteo | Retargeting | BTO 2016 | Daniele Beccari
Criteo | Retargeting | BTO 2016 | Daniele BeccariBTO Educational
 
Large scale-ctr-prediction lessons-learned-florian-hartl
Large scale-ctr-prediction lessons-learned-florian-hartlLarge scale-ctr-prediction lessons-learned-florian-hartl
Large scale-ctr-prediction lessons-learned-florian-hartlPyData
 
Digital Servicing Using Artificial Intelligence
Digital Servicing Using Artificial IntelligenceDigital Servicing Using Artificial Intelligence
Digital Servicing Using Artificial IntelligenceRené Werner
 
VertaMedia Videonuze presentation
VertaMedia Videonuze presentationVertaMedia Videonuze presentation
VertaMedia Videonuze presentationAlex Bornyakov
 
Ppc keywords discovery search labs sao paulo 2010 pavel dolezal
Ppc keywords discovery   search labs sao paulo 2010 pavel dolezalPpc keywords discovery   search labs sao paulo 2010 pavel dolezal
Ppc keywords discovery search labs sao paulo 2010 pavel dolezalPavel Dolezal
 
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...Amazon Web Services
 
How to Quantify the Value of Kafka in Your Organization
How to Quantify the Value of Kafka in Your Organization How to Quantify the Value of Kafka in Your Organization
How to Quantify the Value of Kafka in Your Organization confluent
 
Innovation, Cloud Powered - Dr Werner Vogels
Innovation, Cloud Powered - Dr Werner VogelsInnovation, Cloud Powered - Dr Werner Vogels
Innovation, Cloud Powered - Dr Werner VogelsAmazon Web Services
 
Service Virtualization - Next Gen Testing Conference Singapore 2013
Service Virtualization - Next Gen Testing Conference Singapore 2013Service Virtualization - Next Gen Testing Conference Singapore 2013
Service Virtualization - Next Gen Testing Conference Singapore 2013Min Fang
 
Ed Snelson. Counterfactual Analysis
Ed Snelson. Counterfactual AnalysisEd Snelson. Counterfactual Analysis
Ed Snelson. Counterfactual AnalysisVolha Banadyseva
 
Big Data and Analytics Innovation Summit
Big Data and Analytics Innovation SummitBig Data and Analytics Innovation Summit
Big Data and Analytics Innovation SummitMartin Yan
 
Bridge Your Kafka Streams to Azure Webinar
Bridge Your Kafka Streams to Azure WebinarBridge Your Kafka Streams to Azure Webinar
Bridge Your Kafka Streams to Azure Webinarconfluent
 
Case Study Com Score
Case Study Com ScoreCase Study Com Score
Case Study Com ScoreFM Signal
 
Flink Forward San Francisco 2018: David Reniz & Dahyr Vergara - "Real-time m...
Flink Forward San Francisco 2018:  David Reniz & Dahyr Vergara - "Real-time m...Flink Forward San Francisco 2018:  David Reniz & Dahyr Vergara - "Real-time m...
Flink Forward San Francisco 2018: David Reniz & Dahyr Vergara - "Real-time m...Flink Forward
 

Similar to Scipy_v1_reprint (20)

Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scaleQcon London 2017 -  Architecture overhaul - Ad serving @ Spotify scale
Qcon London 2017 - Architecture overhaul - Ad serving @ Spotify scale
 
Vpon - 廣告效果導向為基礎的行動廣告系統
Vpon - 廣告效果導向為基礎的行動廣告系統Vpon - 廣告效果導向為基礎的行動廣告系統
Vpon - 廣告效果導向為基礎的行動廣告系統
 
AWS Activate Webinar - Growing on AWS
AWS Activate Webinar - Growing on AWSAWS Activate Webinar - Growing on AWS
AWS Activate Webinar - Growing on AWS
 
20141209 meetup hassan
20141209 meetup hassan20141209 meetup hassan
20141209 meetup hassan
 
Leverage Real-Time Purchase Intent to Boost Sales & Customer Growth
Leverage Real-Time Purchase Intent to Boost Sales & Customer GrowthLeverage Real-Time Purchase Intent to Boost Sales & Customer Growth
Leverage Real-Time Purchase Intent to Boost Sales & Customer Growth
 
Criteo | Retargeting | BTO 2016 | Daniele Beccari
Criteo | Retargeting | BTO 2016 | Daniele BeccariCriteo | Retargeting | BTO 2016 | Daniele Beccari
Criteo | Retargeting | BTO 2016 | Daniele Beccari
 
Large scale-ctr-prediction lessons-learned-florian-hartl
Large scale-ctr-prediction lessons-learned-florian-hartlLarge scale-ctr-prediction lessons-learned-florian-hartl
Large scale-ctr-prediction lessons-learned-florian-hartl
 
Digital Servicing Using Artificial Intelligence
Digital Servicing Using Artificial IntelligenceDigital Servicing Using Artificial Intelligence
Digital Servicing Using Artificial Intelligence
 
VertaMedia Videonuze presentation
VertaMedia Videonuze presentationVertaMedia Videonuze presentation
VertaMedia Videonuze presentation
 
Ppc keywords discovery search labs sao paulo 2010 pavel dolezal
Ppc keywords discovery   search labs sao paulo 2010 pavel dolezalPpc keywords discovery   search labs sao paulo 2010 pavel dolezal
Ppc keywords discovery search labs sao paulo 2010 pavel dolezal
 
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...
Slashing Big Data Complexity: How Comcast X1 Syndicates Streaming Analytics w...
 
How to Quantify the Value of Kafka in Your Organization
How to Quantify the Value of Kafka in Your Organization How to Quantify the Value of Kafka in Your Organization
How to Quantify the Value of Kafka in Your Organization
 
No Refresh Needed
No Refresh NeededNo Refresh Needed
No Refresh Needed
 
Innovation, Cloud Powered - Dr Werner Vogels
Innovation, Cloud Powered - Dr Werner VogelsInnovation, Cloud Powered - Dr Werner Vogels
Innovation, Cloud Powered - Dr Werner Vogels
 
Service Virtualization - Next Gen Testing Conference Singapore 2013
Service Virtualization - Next Gen Testing Conference Singapore 2013Service Virtualization - Next Gen Testing Conference Singapore 2013
Service Virtualization - Next Gen Testing Conference Singapore 2013
 
Ed Snelson. Counterfactual Analysis
Ed Snelson. Counterfactual AnalysisEd Snelson. Counterfactual Analysis
Ed Snelson. Counterfactual Analysis
 
Big Data and Analytics Innovation Summit
Big Data and Analytics Innovation SummitBig Data and Analytics Innovation Summit
Big Data and Analytics Innovation Summit
 
Bridge Your Kafka Streams to Azure Webinar
Bridge Your Kafka Streams to Azure WebinarBridge Your Kafka Streams to Azure Webinar
Bridge Your Kafka Streams to Azure Webinar
 
Case Study Com Score
Case Study Com ScoreCase Study Com Score
Case Study Com Score
 
Flink Forward San Francisco 2018: David Reniz & Dahyr Vergara - "Real-time m...
Flink Forward San Francisco 2018:  David Reniz & Dahyr Vergara - "Real-time m...Flink Forward San Francisco 2018:  David Reniz & Dahyr Vergara - "Real-time m...
Flink Forward San Francisco 2018: David Reniz & Dahyr Vergara - "Real-time m...
 

Scipy_v1_reprint

  • 1. SciPy and Real-time Big Data for Site Optimization Pyleus Message Processor Bolt Pyleus Event Worker Bolt Pyleus SciPy Optimizer Bolt Pyleus Update Messenger Bolt SciPy Bayesian Bandit Spout Spout Application State Visitors to Bankrate.com Impressions and Clicks Improve User Experience For more info, contact: Winnie.Cheng@bankrate.com Which Variation to show Bankrate.com Data Science and Engineering Team Example: Pick better story headlines Objective for Site Optimization: Enable fast and cost-efficient ways of testing new designs to improve user experience Algorithmically decide which of two headlines to show user to maximize click-thru-rate (CTR) Computation Framework with Kafka-Storm Simulation Results With more data, algorithm becomes more confident of estimated CTR for each variation Bayesian Multi-Armed Bandit algorithm on Storm Topology decides how often to show each variation by analyzing impressions and clicks Iteration: 100 W1: 56.37% W2: 42.63% Iteration: 1000 W1: 9.82% W2: 90.18% Iteration: 2000 behavior reversal W1: 64.06% W2: 39.94% Iteration: 3000 W1: 94.84% W2: 5.16% Iteration: 4000 W1: 97.12% W2: 2.88%