Create small exercise for using dbt python model on the given dataset

Meeting notes:
before ml
install docker
clone the repo
run docker-compose (download X GB of data / usb disk)
sql ide required

intro motivation - ML = predict registered customers

describe briefly jinja

dbt_project - make consistent dbt_demo / dbt_workshop
leave out/work into more detail why to use identifier

test source and dbt test itx

Testing notes:
Notes:

Env setup

3. be in correct folder: docker exec -w /project/dbt_demo/ -it dbt-demo-dbt-1 /bin/bash
4. run dbt --version where you should see 1.2.0

Project structure

models/staging - add note from core
models/marts/core - core data model with dimensions and facts
models/marts/intermediate - combine two data sources form staging
run dbt debug you should see All checks passed!
run dbt list you should see

There are 3 unused configuration paths:
- models.dbt_workshop.staging
- models.dbt_workshop.marts
- models.dbt_workshop.marts.core

Data sources

name = jaffle_shop
dbt seeds under dbt schema not dbt_seeds
add _src_ga schema: dbt

Staging models

I would enrich by number of orders so it's clear that staging should be creating join on top of base models

Documentation & tests

we have created on src, not stg instead of stg

defining description - move after test

add subfolder where to put it
run docs command?

exercise

split ga and jaffle
add stg ga

snapshots

docker cmd to copy docker cp update.sql dbt-demo-database-1:/update.sql instead of docker cp update.sql dbt-workshop-database-1:/update.sql

falai

labels into seeds
change base, stg, dim
create _ml.yml and _src_ml.yml
run fal ai
change dim_customers to contain flag

dbt_ml.customer_registration_predictio -> dbt_marts.customer_registration_predictio