Evidence-Based Findings
Key observations derived from the filtered UCDP event data, monthly aggregation, annual analysis, and geographic structure used in this project.
Recorded Events
37,892
filtered UCDP events
Reported Deaths
372,593
battle-related deaths
Peak Complete Year
2024
102,120 reported deaths
Civilian Share
6.97%
25,985 recorded deaths
Core Observations
What the Dataset Shows
Six observations derived from the project's event-level, monthly, annual, and demographic analysis.
Finding 01
Reported Fatalities Vary Substantially Across Months
The filtered UCDP dataset contains 37,892 recorded events and 372,593 reported battle-related deaths. After aggregation by month, the data shows substantial variation in reported fatalities across the 2014–2025 coverage period rather than a uniform level of deaths.
Finding 02
March 2022 Recorded the Highest Monthly Total
The highest monthly total in the dataset occurred in March 2022, with 27,111 reported deaths. Monthly aggregation makes it possible to identify these high-volume periods while maintaining a consistent time series for comparison.
Finding 03
2024 Had the Highest Complete-Year Total
Among complete calendar years in the dataset, 2024 recorded 102,120 deaths. This compares with 99,473 in 2022, 76,254 in 2023, and 94,741 in 2025. The 2014 data begins in March and is therefore treated as a partial year rather than directly compared with complete years.
Finding 04
Deaths per Recorded Event Vary by Year
The dataset records different levels of reported deaths per event across years. In 2022, the figure was 13.68 deaths per recorded event, compared with 7.06 in 2023, 8.66 in 2024, and 11.81 in 2025. This provides an additional measure of the intensity of recorded events beyond total deaths alone.
Finding 05
Civilian Share Was Highest in 2022
Civilian deaths account for 25,985 deaths, or 6.97% of the total recorded deaths in the dataset. The civilian share was 18.26% in 2022, followed by 3.09% in 2023, 2.55% in 2024, and 3.01% in 2025. These figures describe the composition of deaths recorded in the dataset.
Finding 06
Event-Level Records Enable Geographic Analysis
The event-level structure preserves geographic fields such as administrative areas and coordinates. This allows the project to examine where recorded events occurred in addition to analyzing changes in reported fatalities over time.
What the Evidence Supports
Reading the Results
The strongest conclusions supported by this analysis concern temporal variation, annual totals, event intensity, and civilian composition. The dataset shows substantial differences in recorded fatalities across months and years.
The monthly analysis identifies March 2022 as the highest recorded month, while the annual analysis identifies 2024 as the highest complete calendar year in the available data.
These findings describe the patterns present in the available dataset. They should not be interpreted as a definitive accounting of all fatalities or as an explanation of why particular patterns occurred.
Interactive Analysis
Explore the Evidence Yourself
Use the interactive Tableau dashboard to examine monthly reported fatalities, yearly patterns, and available geographic views directly.
Open AnalysisData Context
Important Note
The 37,892 figure refers to filtered UCDP event records used in this project; it is not itself a count of fatalities. The corresponding recorded battle-related death total is 372,593.
The Russia-Ukraine dataset covers March 2014 through December 2025. The 2014 coverage is partial, so its annual total should not be directly compared with complete calendar years without accounting for the difference in coverage.
Reported casualty figures may be incomplete or revised as the underlying source datasets are updated. Dataset coverage, definitions, and reporting methodologies can also affect the resulting figures.
This project focuses on quantitative analysis, trends, and visualization. The findings intentionally avoid political interpretation or causal conclusions that cannot be established from the available data alone.