World Bank Data

World Bank Data

Query World Bank Open Data API for development indicators by country and year.

API Base

https://api.worldbank.org/v2/

All requests should include format=json. Responses return [metadata, data].

Country Indicators

# GDP (current US$) for a country
curl -s "https://api.worldbank.org/v2/country/US/indicator/NY.GDP.MKTP.CD?format=json&date=2015:2023&per_page=50"

# Multiple countries (semicolon-separated)
curl -s "https://api.worldbank.org/v2/country/US;CN;IN/indicator/NY.GDP.MKTP.CD?format=json&date=2020:2023"

Common Indicator Codes

Country Metadata

# All countries
curl -s "https://api.worldbank.org/v2/country?format=json&per_page=300"

# Countries by income level (HIC, UMC, LMC, LIC)
curl -s "https://api.worldbank.org/v2/country?format=json&incomeLevel=LIC&per_page=100"

Topic and Source Listings

# List all topics (education, health, etc.)
curl -s "https://api.worldbank.org/v2/topic?format=json"

# Indicators under a topic (topic 4 = Education)
curl -s "https://api.worldbank.org/v2/topic/4/indicator?format=json&per_page=100"

Pagination and Date Ranges

# Pagination: page= and per_page= (max 32500)
curl -s "https://api.worldbank.org/v2/country/all/indicator/SP.POP.TOTL?format=json&date=2023&per_page=300&page=1"

# Date formats: date=2023 (single), date=2015:2023 (range), date=2015;2018;2023 (list)
# Most recent value only
curl -s "https://api.worldbank.org/v2/country/US/indicator/NY.GDP.MKTP.CD?format=json&mrnev=1"

Response Parsing

curl -s "https://api.worldbank.org/v2/country/US;CN;IN/indicator/NY.GDP.PCAP.CD?format=json&date=2020:2023&per_page=100" | python3 -c "
import sys, json
data = json.load(sys.stdin)
for entry in sorted(data[1], key=lambda x: (x['country']['value'], x['date'])):
    if entry['value'] is not None:
        print(f\"{entry['country']['value']} ({entry['date']}): \${entry['value']:,.0f}\")
"

Best Practices

1. Always append format=json to get JSON instead of XML.
2. Use mrnev=1 for the most recent non-null value.
3. Check [0]['total'] in the response metadata for total result count.
4. Country codes follow ISO 3166-1 alpha-2 (US, CN, IN, BR, etc.).
5. Use per_page=300 for country-level queries to get all in one request.
6. Null values indicate missing data for that year; filter them in output.