Data Lakehouse Symposium | Day 1 | Part 1
Download as PPTX, PDF1 like1,837 views
The document discusses the concept of a data lakehouse, highlighting the integration of structured, textual, and analog/IOT data. It emphasizes the importance of common identifiers and universal connectors for meaningful analytics across different data types, ultimately aiming to improve healthcare and manufacturing outcomes through effective data analysis. The presentation outlines the challenges of managing diverse data formats and the potential for data-driven insights to enhance quality of life.
Data Lakehouse Symposium | Day 1 | Part 1
- 1. A presentation by W H Inmon DATA LAKEHOUSE – THE BASIC ELEMENTS
- 2. All data in the corporation Structured data Textual data Analog/IoT data
- 3. Structured data Textual data Analog/IoT data Each of the different types of data have their own unique characteristics
- 4. Structured data Usually transaction based record key attribute index Bank transactions Point of sale Telephone call Payments made Payments received …………………….
- 5. Structured data The same record type is repeated Each record has different contents
- 6. Textual data Medical records Contracts Internet Call centers Warranty claims Insurance claims Email ……………….. Text is found everywhere English Spanish Portuguese French Mandarin Korean German Formal language Slang Acronyms …………….. Voice Written Internet Video ………………..
- 7. Textual data Textual ETL taxonomies Text is transformed Into a structured format
- 8. Analog/IoT Machine generated Drones Electric eye Temperature gauge Speed Mechanical Telemetry …………………….
- 9. Analog/IoT telemetry Date Sept 2, 2021 Time 11:21 am Location from Denver Location to Co Spgs Elevation 786 792 812 854 901 978 1012 1256 1469 1672 2018 2259 2871 …….. Speed 0 35 79 124 197 276 367 416 521 702 835 915 ….. Telemetry data is generated as the rocket is launched and is measured throughout the flight
- 10. Analog/IoT The data lake is created by throwing data all the data into the lake Textual data Structured data
- 11. Analog/IoT Soon the data lake turned into a swamp
- 12. Analog/IoT The data swamp was not good for anyone….
- 13. Analog/IoT The data lake needs to be turned into a lakehouse
- 14. Analog/IoT All this education and 95% of my job is being a data garbageman Data scientist
- 15. Analog/IoT Data scientist Ah, that’s more like it infrastructure
- 16. Analog/IoT Machine generated Time – 0912 Time – 0916 Time – 1002 Time – 1008 Time – 1017 ……………. Basic, raw measurements High probability High performance Low probability Bulk storage Analog/IoT data is often segmented
- 17. High probability High performance Low probability Bulk storage Date of launch Ultimate speed Ultimate height Final landing point Second by second measurements
- 18. Structured data Textual data Analog/IoT data Relative volumes of data in each sector
- 19. Structured data Textual data Analog/IoT data Business value and the volumes of data
- 20. Structured data Textual data Analog/IoT data Relational format Raw data format From a format standpoint, the structured and the textual environments are very different from the analog/IoT environment Format compatibility
- 21. Structured data Textual data Analog/IoT data Key compatibility – very unintegrated Content compatibility
- 22. Structured data Textual data Analog/IoT data In order to do analytics, there must be some common data on which to do a comparison Without common data it is very difficult to do a meaningful comparison
- 23. Structured data Textual data Analog/IoT data The problem is that there may be no obvious, easy way to isolate common identifiers
- 24. Structured data Textual data Analog/IoT data Fortunately there are such things as universal common connectors
- 25. Structured data Textual data Analog/IoT data Universal common connectors exist regardless of the way that data has been collected
- 26. Structured data Textual data Analog/IoT data Universal common connector for anything geography time dollar amount General common connectors
- 27. Structured data Textual data Analog/IoT data Universal common connector for humans gender age race Common connectors for humans
- 28. Structured data Textual data Analog/IoT data Universal common connector for physical objects weight color cost size shape Common connectors for objects
- 29. SOME EXAMPLES Universal common connector
- 30. Healthcare – outcomes analysis
- 31. Did the medicine work? Did the vaccination work? Did the operation have the right effect? Outcome analysis
- 32. Structured data Textual data Analog/IoT data Prolia Estrogen Vitamin D Algaecal Calcitonin Sales of - Doctor’s notes tests diagnosis procedure medication history …………… X rays date location patient age examination results
- 33. Structured data Textual data Analog/IoT data What medicines have been purchased What medicines have been prescribed and/or discussed with doctors By state By age By gender By state By age By gender What outcomes have been achieved By state By age By gender
- 34. What medicines have been purchased By state By age By gender What medicines have been prescribed and/or discussed with doctors By state By age By gender What outcomes have been achieved By state By age By gender Analyses – how does treatment in Utah vary from treatment in Oregon? is Prolia more effective than estrogen? when patients are treated with Algaecal, what other side effects are noticed? do women have better results than men? how much does age affect – the types of treatment for osteoporosis the effectiveness of treatment whether men react differently than women
- 35. What medicines have been purchased By state By age By gender What medicines have been prescribed and/or discussed with doctors By state By age By gender What outcomes have been achieved By state By age By gender When you have both treatment and outcome data together, you can answer – for the first time – important questions about treatment, medication, dosage, side, effects, demographics of treatment You can match outcome with treatment
- 36. What medicines have been purchased By state By age By gender What medicines have been prescribed and/or discussed with doctors By state By age By gender What outcomes have been achieved By state By age By gender The result is healthier people and longer life and better quality of life
- 37. Manufacturing
- 38. Structured data Textual data Analog/IoT data Sales data unit sold date of sale location of sale customer address Warranty claims unit unit type defect severity in use desc Manufacturing data unit id lot id date of manufacture machine used operator
- 39. Textual data Structured data Analog/IoT data Units sold Date of sale Location of sale Unit id Defect description Date of warranty Unit id Machine used for manufacture Date of manufacture Operator Lot id Manufacture telemetry Unit id Unit id Unit id
- 40. Units sold Date of sale Location of sale Unit id Defect description Date of warranty Unit id Machine used for manufacture Date of manufacture Operator Lot id Manufacture telemetry Analyses – what manufacturing machines are producing defects what manufacturing machines are not producing defects what operators are producing defects what operators are not producing defects what telemetry needs to be adjusted under what conditions are defects created …………………………………………………………
- 41. Units sold Date of sale Location of sale Unit id Defect description Date of warranty Unit id Machine used for manufacture Date of manufacture Operator Lot id Manufacture telemetry With all of this data together and able to be analyzed you can now tell what defects can be corrected and what conditions cause defects to occur. The manufacturing process can be materially improved
- 42. Units sold Date of sale Location of sale Unit id Defect description Date of warranty Unit id Machine used for manufacture Date of manufacture Operator Lot id Manufacture telemetry Now manufacturing can be done efficiently and in a cost effective manner
- 43. With analytics from the data lakehouse, you can improve the lives and livelihood of many people