Gwendolyn Ang
0 followers · 3 following · 904 views
on the atlas — 42
- (9+) Q3 Actuary Renewal1 savers
- (9+) Home1 savers
- (9+) Management Dashboard1 savers
- (9+) Home1 savers
- (9+) Home1 savers
- [MedOps] Live Internal Utilization Dashboard 240814 › Copy of Real-Time Progress Tracker1 savers
- (9+) Home1 savers
- (9+) Home1 savers
- (9+) Home1 savers
- (9+) June 17, 20251 savers
- (9+) Home1 savers
- (9+) Q2 Providers Dashboard1 savers
- (9+) June 4-9, 20251 savers
- (9+) May 30-June 1, 20251 savers
- (9+) MLR View1 savers
- Meet - MLR EDA Dataset Huddle1 savers
- books.ms/main/721E037D384E6F80436AFC1A5E9911B21 savers
- Use Checkboxes in Google Sheets to Toggle Numbers in or out of a calculation - YouTube1 savers
- Vaccine Extract - Google Sheets1 savers
- How data observability fits into the different stages in the data pipeline | by Mikkel Dengsøe | Medium1 savers
- Era of Big Data1 savers
- The key technologies behind SQL Comprehension | dbt Developer Blog1 savers
- The Three Levels of SQL Comprehension: What they are and why you need to know about them | dbt Developer Blog1 savers
- Data modeling techniques for more modularity1 savers
- Data transformation: The foundation of analytics work1 savers
- The Case for the ELT Workflow1 savers
- A Love Letter to ETL Tools1 savers
- Accessing Raw Data1 savers
- Why Does Analytics Engineering Exist?1 savers
- What is Analytics Engineering?1 savers
- Reconsidering Career Optionality - by Erik Torenberg9 savers
- Believe In Others | #193 - Boundless by Paul Millerd2 savers
- crushes are often just misplaced ambition - by Isabel42 savers
- Understanding41 savers
- Home / Twitter22 savers
- Why the Canadian Tech Scene Doesn’t Work20 savers
- Rationality: A-Z9 savers
- Pick Up Limes: Why You're Always Tired, + how food can fix it!8 savers
- Intentionally Making Close Friends - LessWrong5 savers
- Discord3 savers
- recent thoughts on meeting people | by Emily | May, 2022 | Medium3 savers
- Ka Pete: Pasyong mahal ni San Jose2 savers
highlights — 477
Compile list of threads
(9+) Q3 Actuary RenewalAdd the four models as dev sources
(9+) HomeMake SME and Legacy add properly in base_metrics
(9+) HomeObtain count of doctors per speciality per hospital
(9+) HomeRow count after join = 30,243,859
(9+) HomeManagement Dashboard
(9+) HomeList companies in question
(9+) June 17, 2025List companies in question
(9+) June 17, 2025List companies in question
(9+) June 17, 2025Filter things out of main code then remove it :’)
(9+) June 4-9, 2025See Modify and study enrollment statuses task for comparisons
(9+) May 30-June 1, 2025Terminated and hired in full year and terminated during extension period
(9+) MLR ViewTerminated and hired in full year and terminated during extension period
(9+) MLR ViewVariability and Veracity
Era of Big DataVolume, Velocity, and Variety
Era of Big DataIt allows you to take your raw source data and find meaning in it for your end business users; this transformation often takes the form of modular data modeling techniques that encourage standardization, governance, and testing.
Data transformation: The foundation of analytics workwill they have difficulty understanding how your data pipelines work? Is transformation documentation accessible, understandable, and easily explorable by business users?
Data transformation: The foundation of analytics workopen source can be budget-friendly, yet will require more maintenance and skilled technical team members.
Data transformation: The foundation of analytics workThis means that data transformation tools should conform to the practices that allow software engineers to ship faster and more reliable code—practices such as version control, automatic testing, robust documentation, and collaborative working spaces.
Data transformation: The foundation of analytics workMetrics should sit within the transformation layer, abstracting out the possibility of business users writing inaccurate queries or conducting incorrect filtering in their BI tools.
Data transformation: The foundation of analytics workstandardized naming conventions, following SQL best practices, and conforming to consistent testing standards
Data transformation: The foundation of analytics workWe recommend creating a style guide before jumping into the code. This way you can write all of your standards for timezones, data types, column naming, and code comments ahead of time.
Data transformation: The foundation of analytics workdbt and the practice of Analytics engineering allow for the rare situation where doing fun thing is also doing the right thing.
Data modeling techniques for more modularitydata teams were writing complex stored procedures with often confusing dependencies. Or sometimes even manually refreshing queries to calculate metrics for visualizations.
Data transformation: The foundation of analytics workBy automating certain data models, data analysts do not need to repeat the same calculations over and over again within the BI layer.
Data transformation: The foundation of analytics workstandardization, version control, integrated documentation, modularity, and testing,
Data transformation: The foundation of analytics workit’s the baseline for increasing the data quality of your business and creating meaningful data for your end users.
Data transformation: The foundation of analytics workWithout data transformation, analysts would be writing ad hoc queries against raw data sources
Data transformation: The foundation of analytics workAs a data person, that’s all you really need to get by—freedom from the fear of breaking stuff.
The Case for the ELT WorkflowActual pen-to-paper development time was a tiny fraction of that period—most of it was patiently waiting for conditions to be right to make the change.
The Case for the ELT WorkflowIf we could’ve repeated the process, maybe we’d test doing a shutdown of the stored procedure pipeline for a couple days in advance of the actual shutoff, just to see what shadow workflows popped out of the woodwork.
The Case for the ELT Workflowwhat nodes were relatively standalone
The Case for the ELT WorkflowThe key thing was just to communicate dependencies, not to fuss over the tooling we used to map out the flows (at which spreadsheets and DOT graph notation were equally as effective)
The Case for the ELT Workflowas a small data team at a growing medium-size business, we focused on new problems—there wasn’t time to circle back to old problems and repay our tech debt.
The Case for the ELT Workflowwhen we deploy an update to a field definition, the change can apply across all historical data, rather than only newly-created records.
The Case for the ELT Workflowthere was no single place to debug our production transformation workflows
The Case for the ELT WorkflowWe’d always have the raw source data available, and could iterate on how we transformed it after the fact.
The Case for the ELT WorkflowIn an ELT workflow, we could separate the data extraction and transformation steps
The Case for the ELT WorkflowThis finality implies freedom—the freedom of analytics teams to transform data to meet the needs of many downstream users (reporting, ML modeling, operational analytics, exploratory analysis, etc).
A Love Letter to ETL ToolsThe accessibility and affordability of these warehouses meant that, as analytics engineers, we now had the keys to an essentially infinitely-scalable supercomputer on which to run SQL queries.
A Love Letter to ETL ToolsWhen you do analytics in a transactional (OLTP) database, life is full of tradeoffs. They’re great at acting on individual rows extremely quickly, but for large query loads like those required for reporting, row-based databases like MySQL are sloooow.
A Love Letter to ETL ToolsAPI documentation quality varies extremely widely
A Love Letter to ETL Toolsdata ingestion tools like Fivetran, HVR, Stitch (and later on their open-source cousins Airbyte & Singer)
A Love Letter to ETL ToolsETL scripting requires human care and feeding, as API outputs can change consistently.
A Love Letter to ETL Toolswhen this data integration inevitably breaks, who’s responsible for fixing it and when?
A Love Letter to ETL ToolsWe could call reliable data ingestion tools + fast, inexpensive and scalable cloud data warehouses the foundation that make the practice of analytics engineering possible.
A Love Letter to ETL Toolsdepends on having reliably-synced data in the warehouse
Accessing Raw Datadata practitioners clean & transform their datasets in isolation, as part of building their own analytics outputs (reports, notebooks, ML models, etc).
Why Does Analytics Engineering Exist?the analytics engineer is a steward of organizational knowledge, not a researcher answering a specific question.
What is Analytics Engineering?you’ll notice they are fascinated by solving a different class of problems than the other members of the data team.
What is Analytics Engineering?