Workshop Agenda
Data Analysis, Automation & Evidence-Based Decision-Making with AI-Augmented R Workflows
It follows the SDPI × D4D schedule of 24 August 2026. Earlier agenda documents that disagreed with each other are preserved in archive/agendas/. Edit this file when the schedule changes.
Dates: 31 August – 4 September 2026 Timing: Monday–Thursday 09:30–16:30 · Friday 09:30–13:00 (closes before Jummah) Trainer: Prof. Dr. Zahid Asghar, School of Economics, QAU · Organisers: SDPI & D4D Audience: researchers and analysts in social science, health, economics and policy Prerequisite: none — Day 1 assumes no prior R
Breaks each full day: tea 11:15–11:30, lunch 13:00–14:00, tea 15:15–15:30.
Day 1 — Foundations, Data Acquisition & AI Toolkit
Monday 31 August · day page
| Time | Session | Material |
|---|---|---|
| 09:30–10:00 | Opening: roadmap, introductions, returning-vs-new poll | 00-orientation |
| 10:00–11:15 | R and Positron fundamentals; project workflow with here::here(); tidyverse recap |
01-r-fundamentals, 02-rstudio-git-session |
| 11:30–13:00 | Data acquisition: CSV/Excel, haven labels, WDI and HDX APIs, pdftools, rvest |
03-data-import-export, import-alifailan.R |
| 14:00–15:15 | AI toolkit setup; AI-assisted cleaning and recoding, reviewed critically | AI prompt library, clean-dirty-excel.R |
| 15:30–16:30 | AI-assisted wrangling: case_when(), .by=, join_by(), pivots |
04-dplyr-wrangling-floods |
Homework: bring one dataset you regularly work with — Days 2–4 use it.
Day 2 — Indicators, Prices & Visualisation
Tuesday 1 September · day page
| Time | Session | Material |
|---|---|---|
| 09:30–11:15 | Key indicators; survey design with srvyr; tables with gt / gtsummary |
gt-summary-tables.R, hies-income-2024.R |
| 11:30–13:00 | Price & inflation — PBS CPI case study: base-year splicing, YoY/MoM, dispersion, index numbers | 01-wfp-prices-pipeline, 04-index-numbers, tidy-pbs-cpi.R |
| 14:00–15:15 | Static and interactive visualisation; Lorenz curves and the Gini | nhanes-ggplot-basics.R |
| 15:30–16:30 | Animated visualisation with gganimate; AI-assisted plotting |
animated-gif.R, temperature-lahore-delhi.R |
Day 3 — Spatial Analysis & Statistical Modelling
Wednesday 2 September · day page
| Time | Session | Material |
|---|---|---|
| 09:30–11:15 | Linear and logistic regression; svyglm with survey weights; diagnostics |
linear-regression.R, 01-dhs-model-datasets |
| 11:30–13:00 | Spatial data and mapping — WFP food prices; sf, boundary joins, tmap |
pakistan-district-choropleth.R |
| 14:00–15:15 | Interactive maps with leaflet; mapview for exploration |
pakistan-maps-geoboundaries.R |
| 15:30–16:30 | Applied policy modelling; affordability from price and wage series; AI-assisted debugging | survey-health-indicators.R |
Day 4 — Automation & Reproducible Reporting
Thursday 3 September · day page
| Time | Session | Material |
|---|---|---|
| 09:30–11:15 | Quarto essentials; cross-references; HTML, PDF and Word output | 01-quarto-basics, 02-quarto-document-writing |
| 11:30–13:00 | Parameterised reports; looped rendering with purrr::walk() |
03-parameterised-reports |
| 14:00–15:15 | Quarto dashboards: value boxes, tabs, embedded maps | penguin-dashboard.Rmd |
| 15:30–16:30 | AI-assisted dashboard development; publishing with quarto publish gh-pages |
04-odi-report, trade-report/ |
Day 5 — Revision, Presentations & Close
Friday 4 September · 09:30–13:00 · day page
| Time | Session | Material |
|---|---|---|
| 09:30–10:30 | Revision — the week in one hour. Acquire → wrangle → measure → see → model/map → report, plus open Q&A | 01-week-in-review |
| 10:30–10:45 | Tea | — |
| 10:45–12:45 | Group presentations | 02-group-presentation-brief |
| 12:45–13:00 | Evaluation, certificates, closing remarks | — |
Day 5 teaches nothing new. The revision slot is one hour; anything that needs longer goes on the follow-up list rather than into the hour.
Group project
An analysis demonstrating the full arc, presented on Day 5:
- import and clean a dataset — survey, price or administrative
- produce at least one correctly weighted or correctly deflated estimate
- one publication-quality table and one chart
- a model or a map
- delivered as a rendered Quarto report, reproducible from the project folder
Materials provided
The workshop data pack (data/ — PDHS, HIES 2024-25, MICS, WFP prices, PBS CPI, Pakistan shapefiles), every script in this repository, the AI prompt library, and a certificate of completion.