tableone is a package for creating "Table 1" summary statistics for a patient population. It was inspired by the R package of the same name by Yoshida and Bohn.
If you use tableone in your study, please cite the following paper:
Tom J Pollard, Alistair E W Johnson, Jesse D Raffa, Roger G Mark; tableone: An open source Python package for producing summary statistics for research papers, JAMIA Open, https://doi.org/10.1093/jamiaopen/ooy012
For documentation, see: http://tableone.readthedocs.io/en/latest/. An executable demonstration of the package is available as a Jupyter Notebook: https://github.com/tompollard/tableone/blob/master/tableone.ipynb. A paper describing our motivations for creating the package is available at: https://doi.org/10.1093/jamiaopen/ooy012.
While we have tried to use best practices in creating this package, automation of even basic statistical tasks can be unsound if done without supervision. We encourage use of tableone
alongside other methods of descriptive statistics and, in particular, visualization to ensure appropriate data handling.
It is beyond the scope of our documentation to provide detailed guidance on summary statistics, but as a primer we provide some considerations for choosing parameters when creating a summary table at: http://tableone.readthedocs.io/en/latest/bestpractice.html.
Guidance should be sought from a statistician when using tableone
for a research study, especially prior to submitting the study for publication.
To install the package with pip, run:
pip install tableone
To install this package with conda, run:
conda install -c conda-forge tableone
- Import libraries:
from tableone import TableOne
import pandas as pd
- Load sample data into a pandas dataframe:
url="https://raw.githubusercontent.com/tompollard/tableone/master/data/pn2012_demo.csv"
data=pd.read_csv(url)
- Optionally, a list of columns to be included in Table 1:
columns = ['Age', 'SysABP', 'Height', 'Weight', 'ICU', 'death']
- Optionally, a list of columns containing categorical variables:
categorical = ['ICU', 'death']
- Optionally, a categorical variable for stratification, a list of non-normal variables, and a dictionary of alternative labels:
groupby = ['death']
nonnormal = ['Age']
labels={'death': 'mortality'}
- Create an instance of TableOne with the input arguments:
mytable = TableOne(data, columns, categorical, groupby, nonnormal, labels=labels, pval=False)
- Type the name of the instance in an interpreter:
mytable
- ...which prints the following table to screen:
Grouped by mortality:
variable | level | isnull | 0 | 1 |
---|---|---|---|---|
n | 864 | 136 | ||
Age | 0 | 66 [52,78] | 75 [62,83] | |
SysABP | 291 | 115.36 (38.34) | 107.57 (49.43) | |
Height | 475 | 170.33 (23.22) | 168.51 (11.31) | |
Weight | 302 | 83.04 (23.58) | 82.29 (25.40) | |
ICU | CCU | 0 | 137 (15.86) | 25 (18.38) |
CSRU | 194 (22.45) | 8 (5.88) | ||
MICU | 318 (36.81) | 62 (45.59) | ||
SICU | 215 (24.88) | 41 (30.15) | ||
mortality | 0 | 0 | 864 (100.0) | |
1 | 136 (100.0) |
- Tables can be exported to file in various formats, including LaTeX, CSV, and HTML. Files are exported by calling the
to_format
method on the tableone object. For example, mytable can be exported to a CSV named 'mytable.csv' with the following command:
mytable.to_csv('mytable.csv')