showstats: quick and compact summary statistics

showstats quickly produces compact summary statistic tables with vertical orientation.

from showstats import show_stats

show_stats(df)

-Date and datetime columns------------------------------------------------------
 Var. N=100      NA%  Min         Max                     Median                
 date_col        0    1501-01-20  1996-04-09              1755-07-20 00:00:00   
 date_col_2      0    1511-12-06  1999-05-05              1776-03-03 00:00:00   
 datetime_col    0    1501-01-20  1996-04-09 06:29:29     1755-07-20 10:10:59   
                      14:37:46                                                  
 datetime_col_2  0    1511-12-06  1999-05-05 14:12:20     1776-03-03 13:25:50   
                      23:40:13                                                  
-Numerical columns--------------------------------------------------------------
 Var. N=100         NA%  Avg     SD     Min     Max     Median 
 float_mean_2       0    2.0     0.89   -0.36   4.12    2.0    
 float_std_2        0    0.14    2.0    -5.17   4.91    0.14   
 float_min_-7       0    -4.64   0.89   -7.0    -2.51   -4.63  
 float_max_17       0    14.88   0.89   12.51   17.0    14.88  
 float_big          0    1.23E6  0.89   1.23E6  1.23E6  1.23E6 
 float_col          0    0.5     0.29   0.0     0.99    0.5    
 U                  0    0.54    0.26   0.02    0.98    0.57   
 int_col            0    49.5    29.01  0       99      49.5   
 int_with_missings  5    48.32   28.8   0       99      49.0   
 bool_col           26   0.5     0.5    false   true    0.5    
 null_col           100                                        
-Categorical columns------------------------------------------------------------
 Var. N=100       NA%  Uniques  Top 1       Top 2        Top 3        
 str_col          48   5        foo (15%)   ABC (13%)    bar (12%)    
 categorical_col  0    2        Fara (57%)  Car (43%)                 
 enum_col         0    3        best (36%)  worst (35%)  medium (29%) 

# Only one type
show_stats(df, "cat")  # Other are num, time

-Categorical columns------------------------------------------------------------
 Var. N=100       NA%  Uniques  Top 1       Top 2        Top 3        
 str_col          48   5        foo (15%)   ABC (13%)    bar (12%)    
 categorical_col  0    2        Fara (57%)  Car (43%)                 
 enum_col         0    3        best (36%)  worst (35%)  medium (29%) 

# Importing showstats adds the stats namespace
df.select("U", "int_col").stats.show()

-Numerical columns--------------------------------------------------------------
 Var. N=100  NA%  Avg   SD     Min   Max   Median 
 U           0    0.54  0.26   0.02  0.98  0.57   
 int_col     0    49.5  29.01  0     99    49.5   

• Primarily built for polars data frames, showstats converts other inputs.

  • For full compatibility with pandas.DataFrames install via pip install showstats[pandas].

• Heavily inspired by the great R-packages skimr and modelsummary.

• Numbers with many digits are automatically converted to scientific notation.

• Because showstats leverages polars effective parallelism, it`s fast: <1s for a 1,000,000 × 1,000 data frame, on a M1 MacBook