Women's College Basketball Roster Database

Cleaned, standardized rosters for divisions I, II and III, with one stable ID that follows a player across every season and team. Maintained by the Sports Roster Data project at the University of Maryland.

Roster explorer

Filter one season at a time by team, conference, state, position and class year, then download exactly what you filtered.

Follow a player across years

Every season she played, with each team and class year. Transfers show up in the timeline. Open an example, or search a name above.

Geography

Hometowns by state and per capita, international recruiting over time, hotbed cities and in-state rates.

Trends over time

Class-year mix (the COVID eligibility freeze is visible), heights, positions and how rosters have grown.

Updates

Loading…

About and downloads

Every CSV, a column-by-column data dictionary, known caveats and how to credit the project.

About this data & this site

Cleaned NCAA women's college basketball rosters, a stable cross-season player ID, and the aggregates built on them.

What this is

The repository behind this site holds one cleaned roster CSV per season (2020-21 onward) plus a cross-season player identifier, wbb_id, that follows a player across teams and seasons. The original roster information is scraped from team websites and augmented by individual research and editing; team information comes from the NCAA.

This data is a project of the JOUR479X class, Sports Data Analysis & Visualization, at the Philip Merrill College of Journalism at the University of Maryland. Please credit the Sports Roster Data project at the University of Maryland.

Student contributors: Ben Baruch, Michael Charles, Shane Connuck, Ian Decker, Jetson Ku, Tanner Malinowski, Varun Shankar, Rina Torchinsky and Dylan Manfre. This project was supervised and edited by Derek Willis, lecturer in data and computational journalism.

Download the data

All files live in the repository, are rebuilt by the cleaning pipeline, and carry one row per player per season. The roster explorer can also export any filtered view as CSV. This site is rebuilt automatically whenever the underlying data changes.

Data dictionary

The site's roster explorer and downloads use these columns:

wbb_id: Stable cross-season player identifier (wbb-######). The same player keeps her ID across seasons and teams; built in player_ids.Rmd.
ncaa_id: The team's NCAA id — the stable team key. Team names change over the years; ids don't.
team: Team name, standardized against teams.csv.
name: Player name as listed on the team's roster.
jersey: Jersey number, when the roster lists one.
year_clean: Standardized class year (Freshman … Sixth Year). Rosters may mean either academic year or athletic eligibility by this field, and COVID eligibility rules added extra years — read it as the roster presented it.
redshirt: 1 when the roster marks the player as a redshirt.
position_clean: Standardized position; may be a combination such as GUARD/FORWARD.
primary_position: GUARD, FORWARD, CENTER or WING, derived from the position field.
total_inches: Height in total inches, so comparisons are possible. Roster values outside 58–90 inches are treated as parse errors and left blank.
hometown_clean: Hometown as printed on the roster, uppercased. Parsed, not normalized — spelling variants remain.
state_clean: US state or territory abbreviation; blank for hometowns outside the US.
country_clean: Country, standardized against a FIBA-derived list; USA for US hometowns.
hs_clean: Standardized high school, where one was matched (~44% of US player-seasons; see HIGH_SCHOOL_STANDARDIZATION.md).
previous_school_clean: Previous school as listed by the roster — sparse, and the least consistent field in the data.
conference: The team's conference, from teams.csv.
division: NCAA division (I, II, III), plus NAIA where identified.
url: The roster page this row was taken from.

Player pages add:

match confidence: How the cross-season ID was assigned: high or medium confidence for matched players, or a single-season record when the player appears in only one season's data.
seasons: The player's season-by-season timeline — season, team, class year, redshirt flag and roster URL.

Known limitations

Transfer counts are a floor. Cross-season matching is automatic for clear cases; borderline matches wait for human adjudication (wbb_id_review_queue.csv in the repository) and are not yet counted as moves. Counts can only rise.
Season coverage runs through what wbb_rosters_combined.csv contains — currently 2020-21 through 2025-26. The newest season file appears here once the annual cross-season ID pipeline is re-run.
Hometowns are parsed but not normalized ("CHICAGO, ILL" can appear next to "CHICAGO, IL"); city tabulations on the geography page are best-effort.
Class years may reflect academic or athletic standing — the COVID eligibility freeze makes them noisier than usual.
Previous school is filled for only a minority of rows and its formatting varies.
Heights outside 58–90 inches are treated as scrape/parse errors and excluded from aggregates.

Methodology & credits

Cleaning and standardization: cleaning.Rmd — heights, positions, class years, hometowns, high schools and previous schools.
Cross-season player IDs: player_ids.Rmd — the tiers, corroborating evidence and human review queue behind wbb_id.
Known data-quality issues in the source scrapes: DATA_QUALITY_ISSUES.md.
The original scraper: dwillis/wbb.

The site is hand-built HTML/CSS/JS — tables by Tabulator, charts by ECharts, US map geometry derived from US Census Cartographic Boundary files. Light theme only for now. Data and code are released under the MIT License. Spotted an error? File an issue.