About ConflictLens

ConflictLens is an open analytical notebook project examining the structure of recorded conflict violence.

The project combines public data, reproducible analysis and explanatory writing. Its aim is to identify patterns in the recorded distribution of violence while documenting the measurement choices and limitations that shape those patterns.

Analytical scope

ConflictLens currently works at two complementary levels.

The first is the analysis-unit-year level. It is used to study concentration, intensity, diffusion, temporal change, macro-contextual gradients and signals of civilian exposure.

The second is the event level. It provides the finer geographical, temporal and actor-related detail that annual aggregates cannot preserve.

Neither level is sufficient on its own. Aggregates reveal structure; events help recover composition and context.

Source hierarchy

UCDP is the primary source for the project’s headline organized-violence death estimates. ACLED is used as a complementary benchmark rather than a drop-in substitute. V-Dem and World Development Indicators provide political, institutional, demographic and socioeconomic context.

Each source follows its own definitions, coverage rules and revision process. Disagreement between sources is therefore treated as information about measurement, not automatically as an error to be eliminated.

Interpretation principles

ConflictLens follows several public guardrails:

  1. A country-year contributor is not a perpetrator.
  2. A share of recorded deaths is not a moral or legal verdict.
  3. Missing data and observed zeroes are not interchangeable.
  4. Raw counts and population-normalized rates answer different questions.
  5. Aggregate curves must be decomposed before they are interpreted.
  6. Descriptive associations are not treated as causal effects.
  7. Recorded conflict data is not a substitute for historical or political analysis.

The project does not forecast future violence, attribute legal responsibility or create rankings of suffering.

Reproducibility

The analytical workflow is published as a sequence of notebooks covering:

  • source inventory and exploratory profiling;
  • construction of the analysis-unit-year panel;
  • descriptive country-year analysis;
  • article-specific reproduction and validation.

Each published finding should be traceable to an executed notebook, a documented source and an explicit analytical convention.

Raw third-party datasets are not redistributed. Reproduction therefore requires downloading the corresponding source versions from their original providers.

The complete workflow is available in the public repository. See the methodology, data-source documentation and interpretation guardrails.

Project status

ConflictLens is an independent and evolving project created by Ludovic Lafon.

The current work focuses on turning the validated analytical foundation into a series of public articles, while progressively extending the analysis from annual aggregates towards event-level detail.