VerdictPal
Join
GuidesPath

Data notebook researcher

Go from raw data to documented, shareable analysis with clear provenance.

Intermediate
Start now
Learning path

Data notebook researcher

Researchers working with datasets who need reproducible analysis workflows. · Go from raw data to documented, shareable analysis with clear provenance.

Lessons5Time~95 minArtifactA documented Jupyter notebook with data cleaning, analysis, visualizations, and a summary of findings.Levelintermediate
Before you start
  • Basic Python or R familiarity
  • A dataset to analyze
  • Jupyter or Google Colab access
Tools in this guide
NotebookLMClaude
0%0/5 steps
Path artifact

A documented Jupyter notebook with data cleaning, analysis, visualizations, and a summary of findings.

Finish every lesson and you should have: A documented Jupyter notebook with data cleaning, analysis, visualizations, and a summary of findings..

Set up the notebook environment

10 minNotebookLM

Create a Jupyter notebook, import core libraries, and document the dataset source and size.

Pitfall
Skipping the data dictionary — future you won't remember what 'col_x' means.
Verification
The first cell documents the data source, row count, and column names.
→ See playbook step 1 in detail

Profile the raw data

15 minClaude

Run shape, dtypes, missing values, and basic statistics before any transformation.

Pitfall
Dropping rows with missing values without documenting how many and why.
Verification
Missing value counts and data types are documented in a summary cell.
→ See playbook step 2 in detail

Clean and transform

25 minClaude

Write reproducible cleaning steps with clear comments: handle missing values, standardize formats, create derived columns.

Pitfall
Applying transformations in-place without a copy — impossible to debug what changed.
Verification
Each transformation cell has a comment explaining what it does and why.
→ See playbook step 3 in detail

Analyze and visualize

30 minClaude

Create exploratory visualizations and summary statistics that answer the research question.

Pitfall
Overplotting — showing every data point when a summary chart tells the story.
Verification
Every chart has a title, axis labels, and a one-sentence takeaway.
→ See playbook step 4 in detail

Document findings

15 min

Write a summary markdown cell with key findings, limitations, and next steps.

Pitfall
Saving the notebook as 'final_v2_actual.ipynb' — use git and version control instead.
Verification
The notebook can be read top-to-bottom by someone who wasn't involved.
Finish line

You should now have: A documented Jupyter notebook with data cleaning, analysis, visualizations, and a summary of findings..

  • The first cell documents the data source, row count, and column names.
  • Missing value counts and data types are documented in a summary cell.
  • Each transformation cell has a comment explaining what it does and why.
  • Every chart has a title, axis labels, and a one-sentence takeaway.
  • The notebook can be read top-to-bottom by someone who wasn't involved.