洁净室入门：在数据砖上使用洁净室在 BI、ML 等中使用更多更好的数据.pdf

1、Clean Room Primer:Using Clean Rooms on Databricks to Utilize More and Better Data in your BI,ML,and BeyondMatt Karasick;Chief Product Officer,HabuAnil Puliyeril;Senior Architect;HabuDatabricks2023Habu is a data clean room software company that makes collaboration across decentralized data safe,simpl

2、e,scalable,and smart.SFNYBOSEMEALATAMF I N A N C I A LT R A V E LG A M I N GA U T O M O T I V E M E D I A&E N T.R E T A I LC P GMatt KarasickChief Product Officer,HabuAnil PuliyerilSenior Architect,Habu H E A L T H C A R ELet your data be your data,and your partners data be theirs.Let your models be

3、 yours,your partners models be theirs.Habus core architectural principle/promise to its adoptersIf you and your partner create mutually beneficial opportunities wherein you explicitly decide to share or commingle data or models,we give you the tooling to do it under the plain light of day,with contr

4、ol.Market forces are resulting in loss of data signal ushering in the era of decentralized data Deprecation of DataConsumers continue to opt out of third-party tracking and data storingConsumer ActionsThird-party cookie restrictions,changes to mobile ad IDs,and the rise of privacy-first browsers imp

5、act identity-based marketing strategiesBrowser and Operating System RestrictionsAd blockersDo not trackClearing away browsing historyApple ITPFirefox ETPMAID restrictionsRegulations impact advertisers ability to track and store dataPrivacy RegulationsThese closed ecosystems restrict data extraction

6、and tracking,which impacts insights access and marketing planning capabilitiesWalled GardensGDPRLGPDAmazonFacebookGoogleCPRAePrivacyRetail Media&CTVsource:Forrester Research,Inc.Data clean rooms are the vehicle to access decentralized data to fuel collaborative BI and ML use cases Whats a data clean