SlideShare ist ein Scribd-Unternehmen logo
1 von 21
Downloaden Sie, um offline zu lesen
Recommendation Systems
       Vipul Sharma
Eventbrite by the Numbers




                      1.5 million events
                   80 million tickets sold
               $1 billion in gross ticket sales
                  Events in 179 countries
Who am I?

Director of Data Engineering
Studied computer science
Machine Learning, Analytics and Big Data
Spam Detection, Consumer Data
Mining, Infrastructure

linkedin.com/in/vipulsharma3
@vipulsharma
vipul@eventbrite.com
Post Event          Creation




               Event Lifecycle
Organization                        Discovery




                      Sale
Create an Event                 Order Now
                  marketplace
Recommendation - What?

• Mechanism to match users with their needs
  •   Ecommerce – what users should buy.
  •   Content – what users should browse.


• Amazon – Product suggestions
• Netflix – Movie suggestions
•Facebook – Newsfeed
• LinkedIn – People you many know
• Eventbrite – Event Picks for you
Recommendations - Why?
• User Acquisition
   •   Bring users to your service
   •   Build long-term trust
   •   Happy customers are happy advertisers
• User Engagement
   •   Engage users with strategic placements
   •   Build site navigation with various funnels
   •   Expose more inventory to users
• Conversion
   •   Upsell
   •   Convert less popular inventory
• Example Attendee Newsletter
Recommendations – How?


                                                                                 Interest
                                                             Social Graph        Graph
                                                             Your friends like   Your friends
                                        Collaborative        Lady Gaga so        who share
                                        Filtering – Item-    you will like       your interest
                                        Item Similarity      Lady Gaga           in music, tech
                                        You like Godfather   (Facebook, Linke    and movies
                    Collaborative       so you will like     dIn)
                    Filtering – User-
                                        Scarface (Netflix)                       are attending
                    User Similarity
                    People who bought                                            SXSW
                    a camera also                                                (Eventbrite)
                    bought batteries
   Item Hierarchy
                    (Amazon)
   You bought a
   camera so you
   need batteries
   (Amazon)
Reason of Progression?

• User data vs Item data
  •   It was hard to collect user meta data vs item meta
      data
  •   Items < Users
  •   Items are less dynamic than users
• Technology Changes
  •   Social graphs
  •   Big Data
  •   Cloud
  •   Crowd Sourcing
Why Social Graph is not Enough

• Events are social
• Events reflect your interests
• Social graphs are dense
• Interests shift while your graph doesn’t
Determining User Interests

• Ask Users
  •   Keep it frictionless
  •   Explain the benefits
• Learn from User Activity
  •   What they bought, browsed, etc
  •   Maintain a consistent taxonomy
      • Ask publisher
      • Use mathematical models
      • Use crowd sourcing
• Use Facebook
  •   Make sure your taxonomy maps with FB intrest data
Social Graphs

• Implicit Graph - Activity
   •   Connections based on activity
   •   Interests trump relationships
   •   We all create an interest graph
• Explicit Graph - Friends
   •   Friends who do not share your interests
   •   Implicit graph is more active than explicit
   •   Explicit graph does not change with your interests
• Mixed – Activity with Friends
   •   Most powerful
Implicit Social Graph
Mixed Social Graph
Who is similar to me?...
Who is more similar to me?
• A two-step process: Identify clusters (via social graph); use the
interest graph to rank recommendations within that cluster
   •   Is a user more similar to one person in his graph or another?
       •   Preferences of the most similar connection will be ranked highest
   •   Clustering applies detailed data from a single user to a group of
       users who are similar
       •   This eliminates the need to ask each user in that group for detailed data
•Building a Social Graph does the clustering for you
   •   Users do most of the work
   •   They self-select into accurate clusters
•Modeling is another option
   •   Models require that you collect learning data from users– but this
       creates friction
   •   Who is more similar to me?
•Recommendation is a Ranking Problem
Put it all together

Item Taxonomy
User Interest
User Graph/Interest Graph
Ranking
Recommendations
Final Product
Future – Content Discovery

Search
  •   Excellent ability to match user queries with content
  •   Limited understanding of each individual user
  •   Limited understanding of user graph
  •   People place the most trust in content and
      recommendations generated by friends
  •   The social graph will improve search
Reviews
  •   Lack personalization
  •   Trust on Internet < Trust of friends
Future – Content Discovery

Entry Point
  •   More recommendation-based funnels
  •   More interconnected funnels
  •   Friends’ suggestions, similar items, editorial
      picks, popular among similar users, etc
Recommendation Systems
  •   More relevant, with more user data
  •   Finer graphs
Questions?




     See it in action. Download our app:

    eventbrite.com/eventbriteapp
Thank You!
@vipulsharma/ vipul@eventbrite.com

Weitere ähnliche Inhalte

Andere mochten auch

How to Promote an Event
How to Promote an EventHow to Promote an Event
How to Promote an Eventeventupload
 
Eventbrite integration
Eventbrite integrationEventbrite integration
Eventbrite integrationFirstGiving
 
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...Renaud Visage
 
The technical side of Internationalization at Eventbrite or "Sh!t we're going...
The technical side of Internationalization at Eventbrite or "Sh!t we're going...The technical side of Internationalization at Eventbrite or "Sh!t we're going...
The technical side of Internationalization at Eventbrite or "Sh!t we're going...Renaud Visage
 
Eventbrite dataplatform and services - Interest graph based recommendations
Eventbrite dataplatform and services - Interest graph based recommendationsEventbrite dataplatform and services - Interest graph based recommendations
Eventbrite dataplatform and services - Interest graph based recommendationsVipul Sharma
 
Linkedin Events and Eventbrite
Linkedin Events and EventbriteLinkedin Events and Eventbrite
Linkedin Events and EventbriteMatthew Ho
 
How Eventbrite leverages marketing to scale recruitment & fills tough positio...
How Eventbrite leverages marketing to scale recruitment & fills tough positio...How Eventbrite leverages marketing to scale recruitment & fills tough positio...
How Eventbrite leverages marketing to scale recruitment & fills tough positio...LinkedIn Talent Solutions
 
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...Renaud Visage
 
Top 12 Eventbrite add ons for Music and Festival Promoters
Top 12 Eventbrite add ons for Music and Festival PromotersTop 12 Eventbrite add ons for Music and Festival Promoters
Top 12 Eventbrite add ons for Music and Festival PromotersEventbrite UK
 
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...The Business Value of APIs - "Bringing the world together via APIs at Eventbr...
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...Renaud Visage
 

Andere mochten auch (12)

How to Promote an Event
How to Promote an EventHow to Promote an Event
How to Promote an Event
 
Eventbrite integration
Eventbrite integrationEventbrite integration
Eventbrite integration
 
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...
Scaling Eventbrite to $1B - Presented at Dublin Web Summit 2012 by Co-founder...
 
The technical side of Internationalization at Eventbrite or "Sh!t we're going...
The technical side of Internationalization at Eventbrite or "Sh!t we're going...The technical side of Internationalization at Eventbrite or "Sh!t we're going...
The technical side of Internationalization at Eventbrite or "Sh!t we're going...
 
Eventbrite dataplatform and services - Interest graph based recommendations
Eventbrite dataplatform and services - Interest graph based recommendationsEventbrite dataplatform and services - Interest graph based recommendations
Eventbrite dataplatform and services - Interest graph based recommendations
 
Linkedin Events and Eventbrite
Linkedin Events and EventbriteLinkedin Events and Eventbrite
Linkedin Events and Eventbrite
 
How Eventbrite leverages marketing to scale recruitment & fills tough positio...
How Eventbrite leverages marketing to scale recruitment & fills tough positio...How Eventbrite leverages marketing to scale recruitment & fills tough positio...
How Eventbrite leverages marketing to scale recruitment & fills tough positio...
 
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...
From a Monolithic to a Distributed API Architecture
 at Eventbrite - Presente...
 
Mogotix
MogotixMogotix
Mogotix
 
Airbnb and the Hotel Industry
Airbnb and the Hotel Industry Airbnb and the Hotel Industry
Airbnb and the Hotel Industry
 
Top 12 Eventbrite add ons for Music and Festival Promoters
Top 12 Eventbrite add ons for Music and Festival PromotersTop 12 Eventbrite add ons for Music and Festival Promoters
Top 12 Eventbrite add ons for Music and Festival Promoters
 
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...The Business Value of APIs - "Bringing the world together via APIs at Eventbr...
The Business Value of APIs - "Bringing the world together via APIs at Eventbr...
 

Ähnlich wie Eventbrite sxsw

Building Personalized Data Products with Dato
Building Personalized Data Products with DatoBuilding Personalized Data Products with Dato
Building Personalized Data Products with DatoTuri, Inc.
 
Mini-training: Personalization & Recommendation Demystified
Mini-training: Personalization & Recommendation DemystifiedMini-training: Personalization & Recommendation Demystified
Mini-training: Personalization & Recommendation DemystifiedBetclic Everest Group Tech Team
 
Eventbrite Data Platform Talk foir SFDM
Eventbrite Data Platform Talk foir SFDMEventbrite Data Platform Talk foir SFDM
Eventbrite Data Platform Talk foir SFDMVipul Sharma
 
Demystifying Recommendation Systems
Demystifying Recommendation SystemsDemystifying Recommendation Systems
Demystifying Recommendation SystemsRumman Chowdhury
 
Designing recommender system for your application
Designing  recommender system for  your applicationDesigning  recommender system for  your application
Designing recommender system for your application孜羲 顏
 
Data Science for Betabrand
Data Science for BetabrandData Science for Betabrand
Data Science for BetabrandKaren Song
 
Betabrand presentation
Betabrand presentationBetabrand presentation
Betabrand presentationKaren Song
 
Data Science for Betabrand
Data Science for BetabrandData Science for Betabrand
Data Science for BetabrandKaren Song
 
Growth Hacking Basics
Growth Hacking BasicsGrowth Hacking Basics
Growth Hacking BasicsMorgan Brown
 
An introduction to Recommender Systems
An introduction to Recommender SystemsAn introduction to Recommender Systems
An introduction to Recommender SystemsDavid Zibriczky
 
Use of data science in recommendation system
Use of data science in  recommendation systemUse of data science in  recommendation system
Use of data science in recommendation systemAkashPatil334
 
Presentation by purshotam verma
Presentation by purshotam vermaPresentation by purshotam verma
Presentation by purshotam vermaRohit malav
 
Bang for the Buck: Social Media Measurement
Bang for the Buck: Social Media MeasurementBang for the Buck: Social Media Measurement
Bang for the Buck: Social Media MeasurementHeidi Miller
 

Ähnlich wie Eventbrite sxsw (20)

Growth hacking
Growth hackingGrowth hacking
Growth hacking
 
Building Personalized Data Products with Dato
Building Personalized Data Products with DatoBuilding Personalized Data Products with Dato
Building Personalized Data Products with Dato
 
Mini-training: Personalization & Recommendation Demystified
Mini-training: Personalization & Recommendation DemystifiedMini-training: Personalization & Recommendation Demystified
Mini-training: Personalization & Recommendation Demystified
 
Eventbrite Data Platform Talk foir SFDM
Eventbrite Data Platform Talk foir SFDMEventbrite Data Platform Talk foir SFDM
Eventbrite Data Platform Talk foir SFDM
 
Demystifying Recommendation Systems
Demystifying Recommendation SystemsDemystifying Recommendation Systems
Demystifying Recommendation Systems
 
Sina Sohangir Presentation on IWMC 2015
Sina Sohangir Presentation on IWMC 2015Sina Sohangir Presentation on IWMC 2015
Sina Sohangir Presentation on IWMC 2015
 
How to use facebook for business
How to use facebook for businessHow to use facebook for business
How to use facebook for business
 
Designing recommender system for your application
Designing  recommender system for  your applicationDesigning  recommender system for  your application
Designing recommender system for your application
 
Lec7 collaborative filtering
Lec7 collaborative filteringLec7 collaborative filtering
Lec7 collaborative filtering
 
Data Science for Betabrand
Data Science for BetabrandData Science for Betabrand
Data Science for Betabrand
 
Betabrand presentation
Betabrand presentationBetabrand presentation
Betabrand presentation
 
Data Science for Betabrand
Data Science for BetabrandData Science for Betabrand
Data Science for Betabrand
 
Growth Hacking Basics
Growth Hacking BasicsGrowth Hacking Basics
Growth Hacking Basics
 
An introduction to Recommender Systems
An introduction to Recommender SystemsAn introduction to Recommender Systems
An introduction to Recommender Systems
 
Recommender Systems
Recommender SystemsRecommender Systems
Recommender Systems
 
Use of data science in recommendation system
Use of data science in  recommendation systemUse of data science in  recommendation system
Use of data science in recommendation system
 
Presentation by purshotam verma
Presentation by purshotam vermaPresentation by purshotam verma
Presentation by purshotam verma
 
Bang for the Buck: Social Media Measurement
Bang for the Buck: Social Media MeasurementBang for the Buck: Social Media Measurement
Bang for the Buck: Social Media Measurement
 
Ashu Desc
Ashu DescAshu Desc
Ashu Desc
 
Testtting
TestttingTesttting
Testtting
 

Kürzlich hochgeladen

Design pattern talk by Kaya Weers - 2024 (v2)
Design pattern talk by Kaya Weers - 2024 (v2)Design pattern talk by Kaya Weers - 2024 (v2)
Design pattern talk by Kaya Weers - 2024 (v2)Kaya Weers
 
Assure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesAssure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesThousandEyes
 
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...panagenda
 
Accelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessAccelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessWSO2
 
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...itnewsafrica
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...Wes McKinney
 
Kuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialKuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialJoão Esperancinha
 
So einfach geht modernes Roaming fuer Notes und Nomad.pdf
So einfach geht modernes Roaming fuer Notes und Nomad.pdfSo einfach geht modernes Roaming fuer Notes und Nomad.pdf
So einfach geht modernes Roaming fuer Notes und Nomad.pdfpanagenda
 
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...itnewsafrica
 
All These Sophisticated Attacks, Can We Really Detect Them - PDF
All These Sophisticated Attacks, Can We Really Detect Them - PDFAll These Sophisticated Attacks, Can We Really Detect Them - PDF
All These Sophisticated Attacks, Can We Really Detect Them - PDFMichael Gough
 
QCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesQCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesBernd Ruecker
 
Testing tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesTesting tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesKari Kakkonen
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Nikki Chapple
 
Infrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsInfrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsYoss Cohen
 
Bridging Between CAD & GIS: 6 Ways to Automate Your Data Integration
Bridging Between CAD & GIS:  6 Ways to Automate Your Data IntegrationBridging Between CAD & GIS:  6 Ways to Automate Your Data Integration
Bridging Between CAD & GIS: 6 Ways to Automate Your Data Integrationmarketing932765
 
Landscape Catalogue 2024 Australia-1.pdf
Landscape Catalogue 2024 Australia-1.pdfLandscape Catalogue 2024 Australia-1.pdf
Landscape Catalogue 2024 Australia-1.pdfAarwolf Industries LLC
 
A Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersA Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersNicole Novielli
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Nikki Chapple
 
Generative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfGenerative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfIngrid Airi González
 
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...Alkin Tezuysal
 

Kürzlich hochgeladen (20)

Design pattern talk by Kaya Weers - 2024 (v2)
Design pattern talk by Kaya Weers - 2024 (v2)Design pattern talk by Kaya Weers - 2024 (v2)
Design pattern talk by Kaya Weers - 2024 (v2)
 
Assure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesAssure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyes
 
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
 
Accelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessAccelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with Platformless
 
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
 
Kuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialKuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorial
 
So einfach geht modernes Roaming fuer Notes und Nomad.pdf
So einfach geht modernes Roaming fuer Notes und Nomad.pdfSo einfach geht modernes Roaming fuer Notes und Nomad.pdf
So einfach geht modernes Roaming fuer Notes und Nomad.pdf
 
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
 
All These Sophisticated Attacks, Can We Really Detect Them - PDF
All These Sophisticated Attacks, Can We Really Detect Them - PDFAll These Sophisticated Attacks, Can We Really Detect Them - PDF
All These Sophisticated Attacks, Can We Really Detect Them - PDF
 
QCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesQCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architectures
 
Testing tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesTesting tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examples
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
 
Infrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsInfrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platforms
 
Bridging Between CAD & GIS: 6 Ways to Automate Your Data Integration
Bridging Between CAD & GIS:  6 Ways to Automate Your Data IntegrationBridging Between CAD & GIS:  6 Ways to Automate Your Data Integration
Bridging Between CAD & GIS: 6 Ways to Automate Your Data Integration
 
Landscape Catalogue 2024 Australia-1.pdf
Landscape Catalogue 2024 Australia-1.pdfLandscape Catalogue 2024 Australia-1.pdf
Landscape Catalogue 2024 Australia-1.pdf
 
A Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersA Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software Developers
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
 
Generative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfGenerative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdf
 
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...
Unleashing Real-time Insights with ClickHouse_ Navigating the Landscape in 20...
 

Eventbrite sxsw

  • 1. Recommendation Systems Vipul Sharma
  • 2. Eventbrite by the Numbers 1.5 million events 80 million tickets sold $1 billion in gross ticket sales Events in 179 countries
  • 3. Who am I? Director of Data Engineering Studied computer science Machine Learning, Analytics and Big Data Spam Detection, Consumer Data Mining, Infrastructure linkedin.com/in/vipulsharma3 @vipulsharma vipul@eventbrite.com
  • 4. Post Event Creation Event Lifecycle Organization Discovery Sale
  • 5. Create an Event Order Now marketplace
  • 6. Recommendation - What? • Mechanism to match users with their needs • Ecommerce – what users should buy. • Content – what users should browse. • Amazon – Product suggestions • Netflix – Movie suggestions •Facebook – Newsfeed • LinkedIn – People you many know • Eventbrite – Event Picks for you
  • 7. Recommendations - Why? • User Acquisition • Bring users to your service • Build long-term trust • Happy customers are happy advertisers • User Engagement • Engage users with strategic placements • Build site navigation with various funnels • Expose more inventory to users • Conversion • Upsell • Convert less popular inventory • Example Attendee Newsletter
  • 8. Recommendations – How? Interest Social Graph Graph Your friends like Your friends Collaborative Lady Gaga so who share Filtering – Item- you will like your interest Item Similarity Lady Gaga in music, tech You like Godfather (Facebook, Linke and movies Collaborative so you will like dIn) Filtering – User- Scarface (Netflix) are attending User Similarity People who bought SXSW a camera also (Eventbrite) bought batteries Item Hierarchy (Amazon) You bought a camera so you need batteries (Amazon)
  • 9. Reason of Progression? • User data vs Item data • It was hard to collect user meta data vs item meta data • Items < Users • Items are less dynamic than users • Technology Changes • Social graphs • Big Data • Cloud • Crowd Sourcing
  • 10. Why Social Graph is not Enough • Events are social • Events reflect your interests • Social graphs are dense • Interests shift while your graph doesn’t
  • 11. Determining User Interests • Ask Users • Keep it frictionless • Explain the benefits • Learn from User Activity • What they bought, browsed, etc • Maintain a consistent taxonomy • Ask publisher • Use mathematical models • Use crowd sourcing • Use Facebook • Make sure your taxonomy maps with FB intrest data
  • 12. Social Graphs • Implicit Graph - Activity • Connections based on activity • Interests trump relationships • We all create an interest graph • Explicit Graph - Friends • Friends who do not share your interests • Implicit graph is more active than explicit • Explicit graph does not change with your interests • Mixed – Activity with Friends • Most powerful
  • 15. Who is similar to me?... Who is more similar to me? • A two-step process: Identify clusters (via social graph); use the interest graph to rank recommendations within that cluster • Is a user more similar to one person in his graph or another? • Preferences of the most similar connection will be ranked highest • Clustering applies detailed data from a single user to a group of users who are similar • This eliminates the need to ask each user in that group for detailed data •Building a Social Graph does the clustering for you • Users do most of the work • They self-select into accurate clusters •Modeling is another option • Models require that you collect learning data from users– but this creates friction • Who is more similar to me? •Recommendation is a Ranking Problem
  • 16. Put it all together Item Taxonomy User Interest User Graph/Interest Graph Ranking Recommendations
  • 18. Future – Content Discovery Search • Excellent ability to match user queries with content • Limited understanding of each individual user • Limited understanding of user graph • People place the most trust in content and recommendations generated by friends • The social graph will improve search Reviews • Lack personalization • Trust on Internet < Trust of friends
  • 19. Future – Content Discovery Entry Point • More recommendation-based funnels • More interconnected funnels • Friends’ suggestions, similar items, editorial picks, popular among similar users, etc Recommendation Systems • More relevant, with more user data • Finer graphs
  • 20. Questions? See it in action. Download our app: eventbrite.com/eventbriteapp