API 209MATH CAMP 2026
FIELD GUIDE/ DATASETS
SOURCEWORLD BANK WDI
YEARS2000-2022
UNITCOUNTRY-YEAR

FIELD GUIDE 02

A common dataset, four policy paths.

Begin with the same documented country-year file. Then choose a policy question that gives a purpose to the coding work.

FROZEN BASELINE

Start with one shared reference file.

The default dataset is a teaching snapshot built from the World Bank World Development Indicators API. It reduces setup problems while preserving real missingness, revisions, unequal time coverage, and variables used in policy analysis.

File Shape Use
Wide CSV One country-year per row; indicators in columns. Default lessons and analysis.
Long CSV One country-year-indicator per row. Reshaping and indicator comparisons.
Indicator dictionary One indicator per row. Tracks, official names, units, definitions, sources, and retrieval dates.
Inspection script Runnable R source. Check schema, key, years, and missingness.
Protected build script Maintainer-only R source. Rebuild frozen files only with explicit opt-in.
FROZEN TEACHING BASELINE

The 2022 endpoint gives the class one shared baseline. Official historical values can change later. The update lab asks students to detect and document those revisions.

Classification warning. income_level_current and lending_type_current describe World Bank metadata available when the teaching snapshot was assembled. A historical classification requires a country-year source. The project joins countries by iso3c because country names may differ across sources.

COMMON CORE, OPTIONAL PATHS

Run the common example first, then choose a complete track.

Everyone begins with the health example. This gives the room one sample, one model, and common language for checking the work. Each optional track then defines the unit, sample, comparison, and limitation required by its own question. Use the complete scripts in code/tracks/.

  1. 01Health: complete country-years, 2000–2022. Compare logged income and under-five mortality while describing current income classifications honestly.
  2. 02Gender: complete country observations in 2022. Compare female secondary enrollment and adolescent fertility in a cross-section, noting missing coverage and regional context.
  3. 03Access: country trajectories at least ten years long. Compare annual change in electricity and internet access using each country’s actual elapsed years.
  4. 04Climate: countries in the latest common complete year. Compare renewable electricity and carbon intensity while showing per-capita growth; the frozen file’s latest complete common year is 2021.

These questions are descriptive and associational. Causal effects require a separate research design.

NINE INDICATORS

The indicator code is also part of the evidence.

Track Indicator code Plain-language role
Health NY.GDP.PCAP.PP.KD GDP per capita, purchasing-power adjusted.
Health SH.DYN.MORT Under-five deaths per 1,000 live births.
Gender SE.SEC.ENRR.FE Female secondary enrollment, gross percent.
Gender SP.ADO.TFRT Births per 1,000 women ages 15-19.
Access EG.ELC.ACCS.ZS Population with access to electricity.
Access IT.NET.USER.ZS Population using the internet.
Climate EG.ELC.RNEW.ZS Renewable share of electricity output.
Climate EN.GHG.CO2.RT.GDP.PP.KD Carbon intensity of economic activity.
Climate NY.GDP.PCAP.KD.ZG Annual real GDP per capita growth.

Use the downloadable dictionary for the official names, exact units, and source notes.

OPTIONAL EXTENSION

Update one indicator without changing the baseline.

Students who want a larger challenge can supervise an agent-assisted request to the official API. Run Rscript code/04-update-indicator.R SH.DYN.MORT under5_mortality. The script writes a retrieval-dated file under data/updates/. The maintainer build will not overwrite the frozen baseline unless it receives an explicit environment variable.

NON-NEGOTIABLES 
Official source only.
New dated output.
No overwrite of the frozen baseline.
Indicator code and unit preserved.
Retrieval date recorded.
Unique keys tested.
Historical revisions reported separately from new years.