tableone

TableOne

class tableone.TableOne(data: DataFrame, columns: list | None = None, categorical: list | None = None, continuous: list | None = None, groupby: str | None = None, nonnormal: list | None = None, min_max: list | None = None, pval: bool | None = False, pval_adjust: str | None = None, htest_name: bool = False, pval_test_name: bool = False, htest: dict | None = None, isnull: bool | None = None, missing: bool = True, ddof: int = 1, labels: dict | None = None, rename: dict | None = None, sort: bool | str = False, limit: int | dict | None = None, order: dict | None = None, remarks: bool = False, label_suffix: bool = True, decimals: int | dict = 1, smd: bool = False, overall: bool = True, row_percent: bool = False, display_all: bool = False, dip_test: bool = False, normal_test: bool = False, tukey_test: bool = False, pval_threshold: float | None = None, include_null: bool | None = True)[source]

If you use the tableone package, please cite:

Pollard TJ, Johnson AEW, Raffa JD, Mark RG (2018). tableone: An open source Python package for producing summary statistics for research papers. JAMIA Open, Volume 1, Issue 1, 1 July 2018, Pages 26-31. https://doi.org/10.1093/jamiaopen/ooy012

Create an instance of the tableone summary table.

Parameters:

data (pandas DataFrame) – The dataset to be summarised. Rows are observations, columns are variables.
columns (list, optional) – List of columns in the dataset to be included in the final table. Setting the argument to None will include all columns by default.
categorical (list, optional) – List of columns that contain categorical variables. If the argument is set to None (or omitted), we attempt to detect categorical variables. Set to an empty list to indicate explicitly that there are no variables of this type to be included.
continuous (list, optional) – List of columns that contain continuous variables. If the argument is set to None (or omitted), we attempt to detect continuous variables. Set to an empty list to indicate explicitly that there are no variables of this type to be included.
groupby (str, optional) – Optional column for stratifying the final table (default: None).
nonnormal (list, optional) – List of columns that contain non-normal variables (default: None).
min_max (list, optional) – List of variables that should report minimum and maximum, instead of standard deviation (for normal) or Q1-Q3 (for non-normal).
pval (bool, optional) – Display computed P-Values (default: False).
pval_adjust (str, optional) –
Method used to adjust P-Values for multiple testing. The P-values from the unadjusted table (default when pval=True) are adjusted to account for the number of total tests that were performed. These adjustments would be useful when many variables are being screened to assess if their distribution varies by the variable in the groupby argument. For a complete list of methods, see documentation for statsmodels multipletests. Available methods include
```
`None` : no correction applied.
`bonferroni` : one-step correction
`sidak` : one-step correction
`holm-sidak` : step down method using Sidak adjustments
`simes-hochberg` : step-up method (independent)
`hommel` : closed method based on Simes tests (non-negative)
```
htest_name (bool, optional) – Display a column with the names of hypothesis tests (default: False).
htest (dict, optional) – Dictionary of custom hypothesis tests. Keys are variable names and values are functions. Functions must take a list of Numpy Arrays as the input argument and must return a test result. e.g. htest = {‘age’: myfunc}
missing (bool, optional) – Display a count of null values (default: True).
ddof (int, optional) – Degrees of freedom for standard deviation calculations (default: 1).
rename (dict, optional) – Dictionary of alternative names for variables. e.g. rename = {‘sex’:’gender’, ‘trt’:’treatment’}
sort (bool or str, optional) – If True, sort the variables alphabetically. If a string (e.g. ‘P-Value’), sort by the specified column in ascending order. Default (False) retains the sequence specified in the columns argument. Currently the only columns supported are: ‘Missing’, ‘P-Value’, ‘P-Value (adjusted)’, and ‘Test’.
limit (int or dict, optional) – Limit to the top N most frequent categories. If int, apply to all categorical variables. If dict, apply to the key (e.g. {‘sex’: 1}).
order (dict, optional) – Specify an order for categorical variables. Key is the variable, value is a list of values in order. {e.g. ‘sex’: [‘f’, ‘m’, ‘other’]}
label_suffix (bool, optional) – Append summary type (e.g. “mean (SD); median [Q1,Q3], n (%); “) to the row label (default: True).
decimals (int or dict, optional) – Number of decimal places to display. An integer applies the rule to all variables (default: 1). A dictionary (e.g. decimals = {‘age’: 0)) applies the rule per variable, defaulting to 1 place for unspecified variables. For continuous variables, applies to all summary statistics (e.g. mean and standard deviation). For categorical variables, applies to percentage only.
overall (bool, optional) – If True, add an “overall” column to the table. Smd and p-value calculations are performed only using stratified columns.
row_percent (bool, optional) – If True, compute “n (%)” percentages for categorical variables across “groupby” rows rather than columns.
display_all (bool, optional) – If True, set pd. display_options to display all columns and rows. (default: False)
dip_test (bool, optional) – Run Hartigan’s Dip Test for multimodality. If variables are found to have multimodal distributions, a remark will be added below the Table 1. (default: False)
normal_test (bool, optional) – Test the null hypothesis that a sample come from a normal distribution. Uses scipy.stats.normaltest. If variables are found to have non-normal distributions, a remark will be added below the Table 1. (default: False)
tukey_test (bool, optional) – Run Tukey’s test for far outliers. If variables are found to have far outliers, a remark will be added below the Table 1. (default: False)
include_null (bool, optional) – Include None/Null values for categorical variables by treating them as a category level. (default: True)

tableone

Summary of the data (i.e., the “Table 1”).

Type:: dataframe

Examples

>>> df = pd.DataFrame({'size': [1, 2, 60, 1, 1],
...                   'fruit': ['peach', 'orange', 'peach', 'peach',
...                             'orange'],
...                   'tasty': ['yes', 'yes', 'no', 'yes', 'no']})

>>> df
   size   fruit tasty
   1   peach   yes
   2  orange   yes
  60   peach    no
   1   peach   yes
   1  orange    no

>>> TableOne(df, overall=False, groupby="fruit", pval=True)

Grouped by fruit
Missing orange peach P-Value

n 2 3 size, mean (SD) 0 1.5 (0.7) 20.7 (34.1) 0.433 tasty, n (%) no 0 1 (50.0) 1 (33.3) 1.000

yes 1 (50.0) 2 (66.7)

…