V-Dem Scraper [$5/1K💰] | Democracy Indices | 1789-2025 Data

V-Dem Democracy Indices Scraper – Country-Year Democracy Data, 1789–2025

Extract V-Dem v16 democracy indices for 202 countries from 1789 to 2025 as clean country-year rows: electoral democracy (v2x_polyarchy), liberal democracy (v2x_libdem), participatory, deliberative and egalitarian components — plus confidence bounds, standard deviations and any other V-Dem indicator. Filter by country and year, choose wide or long output, export JSON, CSV or Excel. $5 per 1,000 rows, no API key, no registration on v-dem.net.

What Can You Do With V-Dem Data?

• Democracy index time series – build a 1789–2025 panel of v2x_polyarchy or v2x_libdem for every country in one run
• Political science research datasets – country-year panels with upper/lower confidence bounds for regression work
• Regime classification and transitions – pair v2x_regime with year and country to study democratic backsliding and autocratization
• Governance and risk dashboards – feed institutional-quality scores into country-risk models, ESG screens or newsroom charts
• Cross-dataset joins – merge on ISO3 (country_text_id) and year with World Bank, V-Dem components or your own panel
• Teaching and replication – pull the exact v16 numbers behind a published figure, with the dataset version recorded on every row

How It Works

The actor downloads the public V-Dem v16 CSV archive from v-dem.net over plain HTTPS — the 15.6 MB Core file by default, or the 25.6 MB Full + Others release. There is no form, no email gate and no API key: the archive URL is public, so a single streaming request is enough.

That CSV is parsed once, filtered by your country list, year range and variable selection, and reshaped into either wide rows (one per country-year) or long rows (one per country-year-variable). Each row keeps vdem_version, and a run summary plus variable metadata are written to the key-value store so you can see exactly which release and which columns produced the output.

Extract V-Dem Democracy Indices Without The Web Download Form

The official route to V-Dem data is a manual download: open v-dem.net, pick a version, choose Core or Full + Others, accept the licence, download a ZIP and parse a wide CSV whose column names are indicator codes. That is fine once. It is painful when you need a reproducible pull, a scheduled refresh, or a subset for 30 countries.

This actor turns that workflow into an API call: countries, variables, startYear, endYear, dataset. The output is already filtered and reshaped, and every row states its vdem_version, so a dataset built in March and refreshed in September is auditable.

Country-Year Democracy Panel Data For Research And Risk Models

Wide format is the default: one row per country-year with one column per requested variable — the shape regression packages, spreadsheet pivots and BI tools expect. Long format emits one row per country-year-variable with variable, variable_label, value, codelow, codehigh and sd, which is what you want for faceted charts, tidyverse-style analysis or a normalised database table.

Add includeConfidenceBounds to get the Bayesian _codelow / _codehigh pair for each selected variable, and includeStandardDeviation for _sd. Setting includeAllVariables exports every base indicator in the chosen release, which produces very wide rows on Core and is meant for research exports rather than dashboards.

Historical Coverage Back To 1789

startYear defaults to 1789, the first year of V-Dem's historical series, and endYear defaults to 2025, the last year in v16. The onlyContemporary flag keeps only rows from the contemporary coding project (project = 0), which drops the historical country-history rows that predate contemporary coding — useful when you want a comparable modern panel rather than the deep historical series.

Countries can be given as names (Sweden, United States), as three-letter V-Dem codes (SWE, USA, BRA), or as historical names that appear in the archive (Siam, Zaire). Empty countries means every country.

Input – Configure A V-Dem Extract

Field	Type	Required	Default	Notes
countries	array	No	[]	Country names, ISO3/V-Dem codes or historical names. Empty = all countries
variables	array	No	five headline indices	Dropdown indices or any V-Dem variable tag; headline selects the five main indices
includeAllVariables	boolean	No	false	Export every base indicator in the release (very wide rows)
startYear	integer	No	1789	First year, inclusive
endYear	integer	No	2025	Last year, inclusive
onlyContemporary	boolean	No	false	Keep only contemporary-coding rows (project = 0)
includeConfidenceBounds	boolean	No	false	Adds <variable>_codelow and <variable>_codehigh
includeStandardDeviation	boolean	No	false	Adds <variable>_sd
outputFormat	string	No	wide	wide = one row per country-year, long = one row per country-year-variable
dataset	string	No	core	core (15.6 MB) or full (Full + Others, 25.6 MB)
maxItems	integer	No	0	Stop after N items; 0 = no limit. In long format each variable counts as one item

Output – V-Dem Dataset Fields

Wide format (default dataset view):

Field	Type	Description
country_name	string	Country name as published in V-Dem
country_text_id	string	Three-letter V-Dem country code (ISO3-style), the join key
country_id	integer	Numeric V-Dem country identifier
year	integer	Observation year
v2x_polyarchy	number	Electoral democracy index (0–1)
v2x_libdem	number	Liberal democracy index (0–1)
v2x_partipdem	number	Participatory democracy index
v2x_delibdem	number	Deliberative democracy index
v2x_egaldem	number	Egalitarian democracy index
histname	string	Historical polity name for the row
vdem_version	string	Dataset release, e.g. v16

Long format view (dataset_schema.json → long):

Field	Type	Description
country_name	string	Country name
country_text_id	string	Three-letter country code
year	integer	Observation year
variable	string	V-Dem variable code
variable_label	string	Human-readable variable label
value	number	Point estimate
codelow	number	Lower confidence bound, when requested
codehigh	number	Upper confidence bound, when requested
sd	number	Posterior standard deviation, when requested
vdem_version	string	Dataset release

The run also writes a run summary and variable metadata record to the key-value store, and exposes the dataset through the standard Apify API URL.

Example Input – Five Countries, Headline Indices, 2000–2025

{
  "countries": ["Sweden", "Brazil", "USA", "India", "Nigeria"],
  "variables": ["headline"],
  "startYear": 2000,
  "endYear": 2025,
  "outputFormat": "wide",
  "dataset": "core",
  "includeConfidenceBounds": true,
  "maxItems": 500
}

Example Output – One Country-Year Row

{
  "country_name": "Sweden",
  "country_text_id": "SWE",
  "country_id": 47,
  "year": 2025,
  "v2x_polyarchy": 0.889,
  "v2x_libdem": 0.853,
  "v2x_partipdem": 0.635,
  "v2x_delibdem": 0.817,
  "v2x_egaldem": 0.802,
  "v2x_polyarchy_codelow": 0.871,
  "v2x_polyarchy_codehigh": 0.904,
  "histname": "Sweden",
  "vdem_version": "v16"
}

Export To JSON, CSV Or Excel And Schedule Yearly Refreshes

Every run writes to the default dataset, downloadable as JSON, CSV, Excel or XML from the run page, and readable through the Apify API or the apify JavaScript/Python clients. Rows stream in as the archive is parsed, so a full 202-country pull does not need a long timeout.

V-Dem ships one major release per year, so a yearly scheduled run (or a monthly run with endYear fixed) is enough to keep a panel current. Set maxItems to cap spend: at $0.005 per row, 10,000 country-years costs about $50. In long format remember that each variable counts as a separate item.

Related Actors

• World Happiness Report Scraper – country-year happiness and governance scores that join to V-Dem on country and year
• CSLB Scraper – another public-record dataset with structured US state-level rows
• Google Search Scraper – find the papers, DOI references and country reports that cite a given V-Dem figure
• Website Content Crawler – pull V-Dem codebook pages and methodology PDFs into clean text for a research pipeline

FAQ

Why use this actor instead of downloading V-Dem from v-dem.net?

Because the download is manual and the file is wide. v-dem.net gives you a ZIP with every country, every year and hundreds of columns; you still filter, reshape and version it yourself. This actor takes countries, variables and a year range, returns only those rows, reshapes to wide or long, stamps vdem_version on every row and can run on a schedule.

Is there a V-Dem API?

V-Dem does not publish a general REST API for the country-year dataset; access is through the website download and the documentation. This actor is the API-shaped alternative: an Apify run endpoint that accepts JSON input and returns a dataset.

What are alternatives to V-Dem for democracy data?

Freedom House's Freedom in the World, the Economist Intelligence Unit Democracy Index, Polity5 and the World Bank's Worldwide Governance Indicators all measure related concepts. V-Dem is the widest in time coverage (1789–2025) and publishes uncertainty estimates for every point score, which the others generally do not.

What is the difference between Core and Full + Others?

Core (15.6 MB) contains the democracy indices and their component indices. Full + Others (25.6 MB) adds all remaining V-Dem indicators plus the external indicators contributed by partner projects. For headline democracy scores, Core is enough.

How many rows will a run produce?

Wide format: one row per country-year. With countries empty and the default 1789–2025 range across roughly 202 countries, that is tens of thousands of rows — use maxItems or a year range to control it. Long format multiplies by the number of selected variables.

Do I need a proxy or an API key?

No. The archive is served over plain HTTPS from v-dem.net and needs no key. The actor runs on 256 MB without a browser and without a proxy.

Can I use the data commercially?

V-Dem v16 is published under CC BY-SA 4.0. Attribution is required, and derivative datasets must carry the same licence. Check the licence text on v-dem.net for your specific use; this README is not legal advice.

SEO Keywords

v-dem dataset, v-dem scraper, v-dem api, democracy index data, democracy indices dataset, v-dem v16 download, country-year panel data, democracy scores by country, v2x_polyarchy data, electoral democracy index, liberal democracy index, regime type data, political science dataset, comparative politics data, governance indicators dataset, democratization research data, political data api, country year democracy data, v-dem csv, political risk data, autocratization data, v-dem alternative

For AI Agents & LLM Apps

Returns country-year rows from the V-Dem v16 archive: identity fields (country_name, country_text_id, country_id, year, histname, vdem_version), the five headline democracy indices by default, plus optional _codelow / _codehigh / _sd columns.

Minimal working input:

{ "countries": ["Sweden"], "startYear": 2020, "endYear": 2025 }

Variant — tidy long rows for charting or a normalised table:

{ "countries": ["BRA", "IND"], "variables": ["v2x_polyarchy", "v2x_libdem"], "outputFormat": "long", "includeConfidenceBounds": true }

Output fields: country_name, country_text_id, country_id, year, histname, vdem_version plus one column per requested variable (wide), or variable, variable_label, value, codelow, codehigh, sd (long).

Behaviors an agent should know:

• countries empty means all countries — always pass it explicitly to avoid a very large run.
• includeAllVariables: true overrides variables and produces extremely wide rows; do not combine it with maxItems assumptions.
• maxItems: 0 (default) means no limit; in long format each variable counts as one item, so a low cap truncates variables, not just years.
• onlyContemporary: true drops historical rows; leave it false for long time series.
• dataset selects the archive (core 15.6 MB, full 25.6 MB); full does not change the output columns unless includeAllVariables is set.

Billing: pay-per-event at $0.005 per dataset row, plus a small per-GB actor-start event. 1,000 rows ≈ $5.

Legal & Compliance Disclaimer

This actor is an independent tool. It is not affiliated with, endorsed by, or sponsored by the V-Dem Institute, the University of Gothenburg, or the V-Dem project.

It accesses only the publicly downloadable V-Dem dataset archive — no login, no paywall bypass, no private endpoints. V-Dem data is published under CC BY-SA 4.0: attribution to the V-Dem Institute and the dataset version (DOI 10.23696/vdemds26 for v16) is required, and derivative datasets must be shared under the same licence. Users are responsible for complying with the V-Dem terms of use and with their own institutional and funder requirements. This section is a factual disclosure, not legal advice.