Michael Roker Data Analytics 📊

Helping others to learn | Udacity Mentor | Data Analyst at Cable Bahamas | Daily tweets on Data | Feel free to DM me!

[Michael Roker Data Analytics 📊 @michaelroker]

Putting your work on GitHub probably is the best decision you can make to build your portfolio.

But what makes a good project stand out?

A good README.

Here are 4 ways to make a good README: 🧵

[Michael Roker Data Analytics 📊 @michaelroker]

Analytics Engineer = translator between pipelines and BI.
Stack: SQL • Git • dbt • Snowflake • Airflow.
Deliverables: clean models, tests, docs, and a dashboard the business trusts. Start small, ship weekly, iterate.

[Michael Roker Data Analytics 📊 @michaelroker]

nice add-ons: S3 for files, GitHub Actions for CI, MWAA to host Airflow, and dbt docs for the team. what would you swap in this stack?

[Michael Roker Data Analytics 📊 @michaelroker]

The “ultimate” data stack for small teams: Airflow + dbt + Snowflake. simple, fast, proven. here’s why:

[Michael Roker Data Analytics 📊 @michaelroker]

guardrails: small warehouses per workload, 60s auto-suspend, query tags for cost. roles per layer + masking policies for PII. simple rules, big savings.

[Michael Roker Data Analytics 📊 @michaelroker]

@via_marketing_ yep, most teams want fewer vendors. but they also want an exit door. my rule: portability first (sql, dbt core, warehouse-native). convenience second.

[Michael Roker Data Analytics 📊 @michaelroker]

Fivetran might by dbt labs.
EL + T under one roof.
Easier? maybe. Lock in? Also maybe.
Curious how this plays out.
https://t.co/agck4s5863

[Michael Roker Data Analytics 📊 @michaelroker]

Dumping your data into S3 doesn’t make it a data lake.

A real data lake has:
- ✅ A catalog (what's in there)
- ✅ Permissions (who can access what)
- ✅ Structure (partitioning, schema)
- ✅ Query tools (Athena, Redshift, etc)

Without these you just have expensive storage.

[Michael Roker Data Analytics 📊 @michaelroker]

Pretty dashboards are great — they should be visually appealing.
But if they aren’t helping your team make better decisions, what’s the point?

A good dashboard doesn’t just look good.
It drives action.
It answers real business questions.

Are your dashboards helping your team make better decisions?

[Michael Roker Data Analytics 📊 @michaelroker]

@ibn_wittig You need to define learn. Learning syntax is one thing, concepts are another. When you get the concepts down (when you would need a join, what type of data structure is needed) to tackle the problem you face takes a much longer time than just syntax.

[Michael Roker Data Analytics 📊 @michaelroker]

@__mharrison__ This is interesting as it really is an annoyance!

[Michael Roker Data Analytics 📊 @michaelroker]

Your first job in analytics is closer than you think. Keep pushing! 💪

[Michael Roker Data Analytics 📊 @michaelroker]

Want to break into Data Analytics?
Here’s your ultimate guide to learning the skills, building a portfolio, and landing your first job as a Data Analyst. 🧵👇

[Michael Roker Data Analytics 📊 @michaelroker]

Upskill with Cloud & Advanced Tools:
As you grow, learn tools that can give you an edge:
•Cloud Platforms: AWS, GCP, or Azure.
•Data Warehousing: Snowflake, Redshift, BigQuery.
•Advanced SQL: Learn window functions, CTEs, and joins.