BIDA study plan

Set up a local analytics environment with a relational database, a Python environment, and a version-controlled project folder

Write five SQL queries that clean a raw table using joins, CASE expressions, and NULL handling

Build a repeatable ingestion script that loads a CSV into the database and logs row counts before and after cleaning

Document the cleaning rules in a short README so the transformation logic is reproducible by someone else

Automate a multi-step transformation pipeline end to end so it runs from a single command

Add validation checks that fail loudly when row counts, null rates, or key uniqueness break expected thresholds

Schedule the pipeline to run on a timer and capture run logs to a file

Refactor duplicated transformation code into reusable functions with docstrings

Select the correct chart type for five different analytical questions and justify each choice in writing

Build a dashboard with at least four coordinated views on a single cleaned dataset

Apply consistent color, labeling, and axis formatting so the dashboard reads without explanation

Add filters that let a viewer slice by at least two dimensions

Critique three published charts and identify the specific encoding or scale decisions that mislead

Sketch a star-schema design for the dashboard's underlying data before building it

Rebuild the dashboard to read from the modeled tables rather than the flat cleaned file

Present the dashboard to one person and record the questions they ask first

Design and build a normalized transactional schema with primary, foreign, and composite keys

Implement slowly changing dimensions to preserve history on at least one attribute

Write queries that answer five business questions against the modeled schema without touching raw tables

Add referential integrity constraints and confirm they reject bad inserts

Write parameterized Python functions that query the modeled database and return typed results

Handle errors and edge cases explicitly rather than letting exceptions surface raw

Write unit tests covering at least three transformation functions, including one failure case

Profile one slow query and reduce its runtime measurably

Implement a reusable function that computes descriptive statistics and flags outliers

Run a hypothesis test on a real question in the dataset and interpret the p-value in plain language

Explain in writing why correlation in the data does not establish causation for one specific finding

Refactor the codebase to remove hardcoded values and move configuration into a settings file

Assemble an end-to-end project that ingests, transforms, models, analyzes, and visualizes in one runnable flow

Write a timed practice set covering all five domains and score it honestly

Review every missed item and write the underlying concept in your own words

Rehearse a five-minute walkthrough of the project as if presenting to a stakeholder