{
  "title": "Counterfactuals teaching data catalogue",
  "catalog_version": 3,
  "dataset_count": 9,
  "datasets": [
    {
      "id": "anscombe",
      "title": "Anscombe's quartet",
      "question": "Can identical summaries hide different data geometry?",
      "analysis_kind": "scatter",
      "row_count": 44,
      "columns": [
        "row",
        "series",
        "x",
        "y"
      ],
      "numeric_columns": [
        "row",
        "x",
        "y"
      ],
      "categorical_columns": [
        "series"
      ],
      "groupable_columns": [
        "series"
      ],
      "default_x": "x",
      "default_y": "y",
      "default_group": "series",
      "json_url": "/data/counterfactuals/anscombe.json",
      "csv_url": "/data/counterfactuals/anscombe.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "6905a130e55c2bef066107982832a71a3d75ed4c6f09e239c99b4723636a7b1e",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "Four constructed x-y datasets with nearly identical conventional summaries and substantially different geometry."
    },
    {
      "id": "insect-sprays",
      "title": "Insect spray counts",
      "question": "How much do group summaries conceal within-spray variation?",
      "analysis_kind": "scatter",
      "row_count": 72,
      "columns": [
        "row",
        "count",
        "spray"
      ],
      "numeric_columns": [
        "row",
        "count"
      ],
      "categorical_columns": [
        "spray"
      ],
      "groupable_columns": [
        "spray"
      ],
      "default_x": "row",
      "default_y": "count",
      "default_group": "spray",
      "json_url": "/data/counterfactuals/insect-sprays.json",
      "csv_url": "/data/counterfactuals/insect-sprays.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "82c8f5b10fcd75b9b96b290338cf23b2410e220fe719eb02ffa63e5bef4960f3",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "Counts of insects in agricultural experimental units treated with six sprays; the documentation does not describe the assignment mechanism."
    },
    {
      "id": "npk-factorial",
      "title": "NPK factorial pea experiment",
      "question": "What changes when a balanced comparison remembers its blocks?",
      "analysis_kind": "npk",
      "row_count": 24,
      "columns": [
        "row",
        "block",
        "N",
        "P",
        "K",
        "yield"
      ],
      "numeric_columns": [
        "row",
        "block",
        "N",
        "P",
        "K",
        "yield"
      ],
      "categorical_columns": [],
      "groupable_columns": [
        "block",
        "N",
        "P",
        "K"
      ],
      "default_x": "block",
      "default_y": "yield",
      "default_group": "N",
      "json_url": "/data/counterfactuals/npk-factorial.json",
      "csv_url": "/data/counterfactuals/npk-factorial.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "002e05973b2f6edc9d4ab76f69ab638f031955939b318179c35ba9285f4622af",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "A 24-plot nitrogen, phosphate, and potassium factorial pea experiment in six blocks. The NPK interaction is confounded by design."
    },
    {
      "id": "orchard-sprays",
      "title": "Orchard spray Latin square",
      "question": "How does spatial design change the comparison?",
      "analysis_kind": "scatter",
      "row_count": 64,
      "columns": [
        "row",
        "row_position",
        "column_position",
        "treatment",
        "decrease"
      ],
      "numeric_columns": [
        "row",
        "row_position",
        "column_position",
        "decrease"
      ],
      "categorical_columns": [
        "treatment"
      ],
      "groupable_columns": [
        "row_position",
        "column_position",
        "treatment"
      ],
      "default_x": "column_position",
      "default_y": "decrease",
      "default_group": "treatment",
      "json_url": "/data/counterfactuals/orchard-sprays.json",
      "csv_url": "/data/counterfactuals/orchard-sprays.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "c5ba25597495ff7bba1ee24f2301bc007b88a8d0e49047c2d61d4161c080adec",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "These rows support description and design practice. They do not, by themselves, identify a causal effect."
    },
    {
      "id": "plant-growth",
      "title": "Plant growth",
      "question": "Which contrast do you mean when there are two treatments?",
      "analysis_kind": "scatter",
      "row_count": 30,
      "columns": [
        "row",
        "group",
        "weight"
      ],
      "numeric_columns": [
        "row",
        "weight"
      ],
      "categorical_columns": [
        "group"
      ],
      "groupable_columns": [
        "group"
      ],
      "default_x": "row",
      "default_y": "weight",
      "default_group": "group",
      "json_url": "/data/counterfactuals/plant-growth.json",
      "csv_url": "/data/counterfactuals/plant-growth.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "30c852e21ebb96c437a3f38cfe32a9ff372a9f1273f713c6ec4bf92c75ccab68",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "An experiment comparing dried plant weight under a control and two treatment conditions; the package documentation does not describe random assignment."
    },
    {
      "id": "sleep-paired",
      "title": "Student's paired sleep data",
      "question": "What do paired observations preserve that two independent means discard?",
      "analysis_kind": "paired",
      "row_count": 10,
      "columns": [
        "row",
        "id",
        "drug_1",
        "drug_2",
        "difference"
      ],
      "numeric_columns": [
        "row",
        "id",
        "drug_1",
        "drug_2",
        "difference"
      ],
      "categorical_columns": [],
      "groupable_columns": [],
      "default_x": "drug_1",
      "default_y": "drug_2",
      "default_group": "id",
      "json_url": "/data/counterfactuals/sleep-paired.json",
      "csv_url": "/data/counterfactuals/sleep-paired.csv",
      "source_url": "https://stat.ethz.ch/R-manual/R-devel/library/datasets/html/sleep.html",
      "source_sha256": "b0780f4363c44529fff9ab3e599ec78b9a48f913e997e4be268a9578c0ae42da",
      "license": "See source documentation",
      "interpretation_boundary": "These rows support description and design practice. They do not, by themselves, identify a causal effect."
    },
    {
      "id": "tooth-growth",
      "title": "Guinea-pig tooth growth",
      "question": "How do dose and supplement interact in the observed outcomes?",
      "analysis_kind": "scatter",
      "row_count": 60,
      "columns": [
        "row",
        "supplement",
        "dose",
        "length"
      ],
      "numeric_columns": [
        "row",
        "dose",
        "length"
      ],
      "categorical_columns": [
        "supplement"
      ],
      "groupable_columns": [
        "supplement",
        "dose"
      ],
      "default_x": "dose",
      "default_y": "length",
      "default_group": "supplement",
      "json_url": "/data/counterfactuals/tooth-growth.json",
      "csv_url": "/data/counterfactuals/tooth-growth.csv",
      "source_url": "https://stat.ethz.ch/R-manual/R-devel/library/datasets/html/ToothGrowth.html",
      "source_sha256": "8e3b3a590815272b050b4ef712b807086ab72c54be726a88badfe702a5ef9b15",
      "license": "See source documentation",
      "interpretation_boundary": "These rows support description and design practice. They do not, by themselves, identify a causal effect."
    },
    {
      "id": "ucb-admissions",
      "title": "UC Berkeley admissions, 1973",
      "question": "How can aggregation reverse the comparison?",
      "analysis_kind": "admissions",
      "row_count": 12,
      "columns": [
        "row",
        "department",
        "gender",
        "admitted",
        "rejected",
        "applications",
        "admission_rate"
      ],
      "numeric_columns": [
        "row",
        "admitted",
        "rejected",
        "applications",
        "admission_rate"
      ],
      "categorical_columns": [
        "department",
        "gender"
      ],
      "groupable_columns": [
        "department",
        "gender"
      ],
      "default_x": "applications",
      "default_y": "admission_rate",
      "default_group": "gender",
      "json_url": "/data/counterfactuals/ucb-admissions.json",
      "csv_url": "/data/counterfactuals/ucb-admissions.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "provenance_level": "archive-member",
      "source_sha256": "b8dc9b4543660c7b596b87938df532394350360976527d344228ee0ed12e45ec",
      "source_member_path": "R-4.6.0/src/library/datasets/data/UCBAdmissions.R",
      "source_member_sha256": "86a5801b122ec1f172098303de35c9df9a8c4c09f2f81dc9f03bfafdde102d5e",
      "source_documentation_url": "https://stat.ethz.ch/R-manual/R-patched/library/datasets/html/UCBAdmissions.html",
      "source_documentation_member_path": "R-4.6.0/src/library/datasets/man/UCBAdmissions.Rd",
      "source_documentation_sha256": "9f80bbfe5c66697c0ab35482567a4e23e472857d293f5ad7b64de7135a3c8814",
      "normalized_source_sha256": "fdf1d803e59e59e7d97161e4c4db5c15465afee15a731b0d1eb1f1381b03e714",
      "license_url": "https://www.r-project.org/Licenses/",
      "rights_statement": "R Core Team's UCBAdmissions documentation states that this dataset member is distributed under GPL 2 or later.",
      "license": "GPL-2.0-or-later",
      "interpretation_boundary": "These aggregate application records establish an aggregation reversal. They do not identify a causal effect of gender or department without a causal model for application, admission, and omitted qualifications."
    },
    {
      "id": "warp-breaks",
      "title": "Warp-break counts",
      "question": "What changes when wool type and tension are examined together?",
      "analysis_kind": "scatter",
      "row_count": 54,
      "columns": [
        "row",
        "breaks",
        "wool",
        "tension"
      ],
      "numeric_columns": [
        "row",
        "breaks"
      ],
      "categorical_columns": [
        "wool",
        "tension"
      ],
      "groupable_columns": [
        "wool",
        "tension"
      ],
      "default_x": "row",
      "default_y": "breaks",
      "default_group": "wool",
      "json_url": "/data/counterfactuals/warp-breaks.json",
      "csv_url": "/data/counterfactuals/warp-breaks.csv",
      "source_url": "https://cran.r-project.org/src/base/R-4/R-4.6.0.tar.gz",
      "source_sha256": "c5507081e169b009fc3bf59f2974f861f5aaa7635e1b897ea57d8f94ed5a778c",
      "license": "GPL-2-or-later",
      "interpretation_boundary": "Break counts for 54 looms, balanced across two wool types and three tension levels; the documentation does not establish a randomized assignment mechanism."
    }
  ]
}