← All courses

Course 2Actuarial Data Wrangling, Relational Schemas & Reporting

Master pandas, relational operations, spatial feature engineering without GUI tools, and programmatic Excel/dashboard generation.

For: Students comfortable with core Python who need to process real-world messy data, query multi-table insurance schemas, and automate stakeholder deliverables.

Lesson 1 Tabular Data Structures with pandas
Series and DataFrames, .loc vs .iloc, types and missing values. Lab: cleaning a messy claims extract.

Read → Slides Notebook ↓
Lesson 2 Temporal Computations & Exposure Calculations
Parsing dates, Timedelta vs DateOffset, leap years and cancellations. Lab: exact earned exposure.

Read → Slides Notebook ↓
Lesson 3 Relational Data Modeling & Multi-Table Joins
Keys, merge types, cardinality, validation. Lab: a six-table analytical base table that reconciles.

Read → Slides Notebook ↓
Lesson 4 Aggregations, Grouping & Classical Actuarial Metrics
groupby, pivot_table, melt; frequency, severity, pure premium, loss ratio. Lab: multi-way segmentation.

Read → Slides Notebook ↓
Lesson 5 Exploratory Data Analysis & Statistical Visualization
matplotlib and seaborn for loss distributions and risk curves; executive formatting; adverse selection.

Read → Slides Notebook ↓
Lesson 6 Programmatic Spatial Data Analysis
GeoPackages, CRS, spatial joins, distance to coast. Lab: territory rating factors on real NC geography.

Read → Slides Notebook ↓
Lesson 7 Automated Excel Engineering with openpyxl
Styled, formula-driven workbooks and automatic formula testing. Lab: the underwriting analytics panel.

Read → Slides Notebook ↓
Lesson 8 · Part A Pipeline Architecture & Configuration
Extract, features, calculate, report; TOML configs, a CLI, logging, validation gates, tests.

Read → Slides Notebook ↓
Lesson 8 · Part B Capstone: End-to-End Rating Engine & Report
One command from database dumps to an Excel report and a generated Reveal.js deck.

Read → Slides Notebook ↓