Longitudinal Dataset of University Student Game Design Documents (GDDs)

Longitudinal Dataset of University Student Game Design Documents (GDDs) This Zenodo record provides two CSV files used in the analyses reported in: Tapscott, Alan; León, Carlos; Pérez, Óliver; Colás, Joaquim (2025). A Longitudinal Look Into University Student Video Game Designs.DiGRA Digital Library (Conference Proceedings of DiGRA 2025: Games at the Crossroads).DOI (paper): https://doi.org/10.26503/dl.v2025i2.2458 The dataset supports a longitudinal view of how student game design documents evolve over time in university video game courses, combining: Git-derived change statistics at the file level across multiple projects and semesters, and A document-feature coding table (counts/indicators) extracted from a set of GDDs. Important: This dataset contains derived measurements and metadata (including repository URLs). It does not include the full text of students’ design documents. Contents gitGDDStats.csvFine-grained, file-level change statistics derived from version control history (Git) across multiple student projects.Each row corresponds to a (project, file, commit timestamp) observation with the number of changed lines. qualiGDDStats.csvA compact table (one row per student project/group) with document-level features and counts (e.g., word counts by section, number of references, presence/size of flowcharts/lists/tables, iteration documentation indicator). Dataset authors The dataset authors are the same as the paper authors: Alan Tapscott Óliver Pérez Carlos León Joaquim Colás Contact Carlos León — [e-mail] Quick summary (computed from the provided CSVs) gitGDDStats.csv Rows: 126,976 Columns: 11 Time span (commit timestamps): 2017-10-13 15:53:34 UTC → 2024-07-01 02:35:16 UTC Unique projects (project_name): 51 Unique repositories (project_url): 55 Unique semesters (semester_id): 11 Unique files (file_name): 28,382 Rows flagged as GDD (is_gdd=True): 1,062 In this export, is_gdd=True corresponds to file_name="README.md". Course distribution (course_id, rows): - proyectos2=72,818, pvli=32,203, dvi=21,955 qualiGDDStats.csv Rows: 40 Columns: 14 Unique groups (Código grupo): 40 Team size (Miembros): min=1, max=5, mean≈3.70 Selected quantitative fields (means, from this file): - Referentes (references): mean≈3.55 - Palabras intro: mean≈634.02 - Palabras funcional: mean≈3153.53 - Palabras plan desarrollo: mean≈955.27 Missingness examples: - Estados flowchart: 17.5% missing - Celdas tabla: 57.5% missing How to interpret the files 1) gitGDDStats.csv (Git-derived file-change events) Each row represents a file-level observation from Git history, with: When the change happened (date) What file changed (file_name) How large the change was (changed_lines) Whether the file was considered part of the GDD (is_gdd) Which semester and course the project belongs to (semester_id, course_id) Derived “relative” time offsets (relative_semester_date, relative_project_date) Project identifiers (project_name, project_url) This table is useful for analyses such as: - Comparing the update frequency of GDD vs. code, - Measuring how documentation evolves across semesters, - Relating documentation work to project timeline markers (relative dates). Notes on key columns changed_lines is a positive integer representing the magnitude of the modification at the file level (as exported in this dataset). is_gdd is a boolean tag. In this export, it is True for README.md entries (used as the GDD carrier in the analyzed repos). relative_semester_date and relative_project_date are stored as strings like "102 days 14:58:21" and can be parsed as timedeltas. Negative values may occur (e.g., commits before an official start marker). 2) qualiGDDStats.csv (Document-level features / counts) Each row corresponds to a student team/project/group and summarizes structural features of the game design document, including: Team metadata: title, group code, team size, genre Counts: number of references and word counts for key sections A compact description of design artifacts (flowcharts, maps, sprites, screenshots, tables, lists, etc.) A binary indicator of whether iterations were documented (Documentación iteraciones) This table is useful for: - Quantifying differences in document structure and content density - Inspecting which artifacts (flowcharts, tables, lists) are favored - Studying how iteration/documentation practices vary between groups Data dictionary gitGDDStats.csv columns Column Type Description date string (datetime) Commit timestamp (parseable as datetime). file_name string Path/name of the file modified in the commit. changed_lines int Magnitude of change for that file in that commit (as exported). is_gdd bool Whether the file is considered a GDD artifact in the analysis. semester_id string Semester identifier (e.g., s1_2021_2022, s2_2022_2023). relative_semester_date string (timedelta) Time offset relative to a semester anchor (see paper/method for anchor definition). relative_project_date string (timedelta) Time offset relative to a project anchor (see paper/method for anchor definition). course_id string Course identifier (categorical code such as pvli, dvi, proyectos2). project_name string Human-readable project label. project_url string (URL) Repository URL used to mine Git history (often GitHub). Unnamed: 0 int Row index from the export process (can be ignored). qualiGDDStats.csv columns Column Type Description Título string Game/project title. Código grupo string Group identifier (one per row). Miembros int Team size (number of members). Género string Genre label(s) provided/assigned. Referentes int Number of references / inspirations recorded. Palabras intro int Word count for the introduction section. Palabras funcional int Word count for the functional design section. Palabras plan desarrollo int Word count for the development plan section. Elementos diseño string Compact encoding of included design artifacts and counts (e.g., flowcharts, lists, tables, screenshots, maps). Contextos de juego float Numeric indicator/count of gameplay contexts (may contain missing values). Estados flowchart float Number of states or nodes in flowcharts (may contain missing values). Elementos lista float Count of list elements (may contain missing values). Celdas tabla float Count of table cells (often missing when no tables were present/counted). Documentación iteraciones string Iteration documentation flag (s=yes, n=no). Reproducible loading examples Python (pandas) import pandas as pd git = pd.read_csv("gitGDDStats.csv") quali = pd.read_csv("qualiGDDStats.csv") # Parse datetimes / timedeltas git["date"] = pd.to_datetime(git["date"], utc=True, errors="coerce") git["relative_semester_date"] = pd.to_timedelta(git["relative_semester_date"], errors="coerce") git["relative_project_date"] = pd.to_timedelta(git["relative_project_date"], errors="coerce") # Example: compare change magnitudes for GDD vs non-GDD git.groupby("is_gdd")["changed_lines"].describe() # Example: per-project total change in README.md (GDD carrier) gdd_changes = git[git["is_gdd"]].groupby("project_name")["changed_lines"].sum().sort_values(ascending=False) # Example: semester-level aggregation semester_gdd = git[git["is_gdd"]].groupby("semester_id")["changed_lines"].sum() R git <- read.csv("gitGDDStats.csv", stringsAsFactors = FALSE) quali <- read.csv("qualiGDDStats.csv", stringsAsFactors = FALSE) git[["date"]] <- as.POSIXct(git[["date"]], tz="UTC") Suggested analysis patterns Depending on your research question, common analysis paths include: Update frequencyCount commits touching README.md (GDD) versus source code files by time window (weekly/biweekly) and compare across semesters. Timeline alignmentUse relative_semester_date and relative_project_date to align all projects onto a comparable timeline.For example, compare whether documentation increases near deadlines. Document structure vs. iterationJoin qualiGDDStats.csv with aggregate Git metrics (per project) to explore relationships between: word counts and revision activity, presence of diagrams/tables and revision patterns, iteration documentation flag (s/n) and update frequency. To merge the two files you typically need a project/group key. In some pipelines, this is done by mapping Código grupo / Título to project_name. If your analysis requires a precise join, consult the workflow described in the associated paper or your internal mapping tables used during the study. Data quality and known caveats Repository URLs may identify individuals.The project_url field often points to GitHub repositories and may include usernames. This can make re-identification possible in combination with other information. Not all fields are complete.In qualiGDDStats.csv, some counts are missing (NaN) when an element was absent or not recorded in the coding process. Long-tailed change sizes.In gitGDDStats.csv, changed_lines has a heavy-tailed distribution. Consider robust statistics (median, percentiles) or log-scaling where appropriate. Export index column.Unnamed: 0 is a leftover index from the export process and can be dropped. Ethics, privacy, and responsible use This dataset is derived from student projects and coursework artifacts. If you reuse or redistribute: - Avoid attempts to identify individuals from repository URLs or commit histories. - Consider removing or hashing project_url if you only need aggregated statistics. - When reporting results, prefer aggregated insights over project-specific callouts unless you have explicit permission. License Select a Zenodo license that matches your intended reuse policy and the permissions you have for redistributing these derived measurements. Common choices for research datasets: - CC BY 4.0 (attribution required), or - CC BY-NC 4.0 (non-commercial), if you need additional restriction. How to cite Cite the dataset (this Zenodo record) Tapscott, A., León, C., Pérez, Ó., & Colás, J. (YEAR). Longitudinal Dataset of University Student Game Design Documents (GDDs) (Version VERSION) [Data set]. Zenodo. DOI: 10.5281/zenodo.XXXXXXX BibTeX (dataset) @dataset{digra_gdd_longitudinal_dataset, author = {Tapscott, Alan and Le{\'o}n, Carlos and P{\'e}rez, {\'O}liver and Col{\'a}s, Joaquim}, title = {Longitudinal Dataset of University Student Game Design Documents (GDDs)}, year = {2025}, publisher = {Zenodo}, version = {1.0}, doi = {10.5281/zenodo.18790769}, url = {https://doi.org/10.5281/zenodo.18790769} } Cite the related paper APA-style > Tapscott, A., León, C., Pérez, Ó., & Colás, J. (2025). A Longitudinal Look Into University Student Video Game Designs. Conference Proceedings of DiGRA 2025: Games at the Crossroads. https://doi.org/10.26503/dl.v2025i2.2458 BibTeX (paper, from DiGRA Digital Library) @Conference{digra2458, title = {A Longitudinal Look Into University Student Video Game Designs}, year = {2025}, author = {Tapscott, Alan and Le{\'o}n, Carlos and P{\'e}rez, {\'O}liver and Col{\'a}s, Joaquim}, publisher = {DiGRA}, address = {Tampere}, howpublished = {\url{https://doi.org/10.26503/dl.v2025i2.2458}}, booktitle = {Conference Proceedings of DiGRA 2025: Games at the Crossroads} } Changelog v1.0.0 — Initial release on Zenodo (two CSV files: gitGDDStats.csv, qualiGDDStats.csv).