WELCOME
USING DOMO TO BUILD A MODERN PLATFORM Madison Logic 3
USING DOMO TO BUILD A MODERN PLATFORM Ajay Sathyanath Akhil Patlolla Big Data Engineer CTO Madison Logic Madison Logic 4
INTRODUCTION TO MADISON LOGIC B2B Multi-Channel Account Based Marketing Technology • Platform Empower Marketers to Identify, Prioritize and Engage Top • Accounts Recent Key Launches: • ML Data Cloud • Journey Acceleration™ • LinkedIn™ Integration • ML Platform™ • 5
AGENDA Problems in Solutioning A Modern Platform Development Requirements Tool Selection Details of Embedding Strategies How We Built the ML Platform Demo Best Practices Questions 6
PROBLEM: TWO DISPARATE STRATEGIES Build a Modern Client Facing Platform: Power Organization’s BI Requirements: • Democratize Analytics and Provide Actionable Insights on • Enable Self Serve Platform 02 01 • Ease of Use and Low Learning • Decrease Development Time Curve • Maintain Metrics Integrity and • Speed and Connectivity to Reduce Copies Varied Databases Require One Tool to Address Both 7
PLATFORM DEVELOPMENT REQUIREMENTS 01 02 03 Flexibility in Embedding Ability and Ease of Ability to Appear Dashboards / Charts Embedding Native 05 04 06 Passing Filters to Cross Chart Interaction Chart Customization Preload Data 8
TOOL SELECTION METHODOLOGY Factors Scoring Mobile Easy Chart Coolness Cross Chart Friendly Building Factor Interaction Connection Stability Speed Embedding to Live DBs 9
EMBED STRATEGIES Embed Page and Stories Embed Cards Pros: Pros: • Cross card interactions within the embedded • Supports adding external components view • Can use global filters • Better visual customizations compared to cards • Parameterized filtering available • Ideal to use for single page application Caveats: Caveats: External components are not supported • • No Cross-Card Interaction Does not look native • • Unable to carry over dashboard properties 10
EMBED STRATEGIES CONTD. Domo Phoenix JS Native Graphing libraries Highcharts / D3js Pros: Pros: • Chart customization • Offload computation to Domo • Apps looks native • Cross card interaction • Well documented • Can make custom charts available on dashboard • Modularized open source code available • App looks native Caveats: Caveats: • Complex to build and maintain • Development pace slower than pages and cards • Cross chart interaction requires coding • Increases application maintenance • Increased development time 11
SOLUTIONING THE ML PLATFORM • Embed Charts with Parameterized Filters, and Phoenix Components Port Communication via PostMessage to Enable Chart Interaction • Okta Authentication and Domo PDP Access Control • Angular 8 Framework • • Combination of Proprietary and Domo Schedulers to Move Data Node.js Back End • • Combination of Domo and Federated Datasets Data Modeling in Snowflake • 12
ML Platform™ Demo
ML PLATFORM™ DEMO 14
ML PLATFORM™ DEMO CONTD. 15
ML PLATFORM™ DEMO CONTD. 16
BEST PRACTICES 1 4 Use SSO to Authenticate and Use Live Queries for Very Large PDP for Access Control Volume Datasets 2 5 Model Data in Warehouse Create Embed Page Dashboard before Ingesting in Domo Use Parameterized Filtering to 3 6 Use Domo CLI to Create Backups Preload Pages/Cards for Improved Performance 17
CONCLUSIONS Design & Speed is Bound by the Embed Strategy Domo Embed is Powerful and Easy to Use Madison Logic Leveraged Domo to Release ML Platform™ in Record Time 18
THANK YOU
Recommend
More recommend