MartinThoma / edapy

Exploratory Data Analysis with Python

edapy is a first resource to analyze a new dataset.

Installation

$ pip install git+https://github.com/MartinThoma/edapy.git

For the pdf part, you also need pdftotext:

$ sudo apt-get install poppler-utils

Usage

$ edapy --help
Usage: edapy [OPTIONS] COMMAND [ARGS]...

  edapy is a tool for exploratory data analysis with Python.

  You can use it to get a first idea what a CSV is about or to get an
  overview over a directory of PDF files.

Options:
  --version  Show the version and exit.
  --help     Show this message and exit.

Commands:
  csv     Analyze CSV files.
  images  Analyze image files.
  pdf     Analyze PDF files.

The workflow is as follows:

• edapy pdf find --path . --output results.csv creates a results.csv for you. This results.csv contains meta data about all PDF files in the path directory.
• edapy csv predict --csv_path my-new.csv --types types.yaml will start / resume a process in which the user is lead through a series of questions. In those questions, the user has to decide which delimiter, quotechar is used and which types the columns have.
• edapy generates a types.yaml file which can be used to load the CSV in other applications with df = edapy.load_csv(csv_path, yaml_path).

Example types.yaml

For the Titanic Dataset, the resulting types.yaml looks as follows:

columns:
- dtype: other
  name: Name
- dtype: int
  name: Parch
- dtype: float
  name: Age
- dtype: other
  name: Ticket
- dtype: float
  name: Fare
- dtype: int
  name: PassengerId
- dtype: other
  name: Cabin
- dtype: other
  name: Embarked
- dtype: int
  name: Pclass
- dtype: int
  name: Survived
- dtype: other
  name: Sex
- dtype: int
  name: SibSp
csv_meta:
  delimiter: ','

A sample run then would look like this:

$ edapy csv predict --types types_titanik.yaml --csv_path train.csv
Number of datapoints: 891
2018-04-16 21:51:56,279 WARNING Column 'Survived' has only 2 different values ([0, 1]). You might want to make it a 'category'
2018-04-16 21:51:56,280 WARNING Column 'Pclass' has only 3 different values ([3, 1, 2]). You might want to make it a 'category'
2018-04-16 21:51:56,281 WARNING Column 'Sex' has only 2 different values (['male', 'female']). You might want to make it a 'category'
2018-04-16 21:51:56,282 WARNING Column 'SibSp' has only 7 different values ([0, 1, 2, 4, 3, 8, 5]). You might want to make it a 'category'
2018-04-16 21:51:56,283 WARNING Column 'Parch' has only 7 different values ([0, 1, 2, 5, 3, 4, 6]). You might want to make it a 'category'
2018-04-16 21:51:56,285 WARNING Column 'Embarked' has only 3 different values (['S', 'C', 'Q']). You might want to make it a 'category'

## Integer Columns
Column name: Non-nan  mean   std   min   25%   50%   75%   max
PassengerId:     891  446.00  257.35     1   224   446   668   891
Survived   :     891  0.38  0.49     0     0     0     1     1
Pclass     :     891  2.31  0.84     1     2     3     3     3
SibSp      :     891  0.52  1.10     0     0     0     1     8
Parch      :     891  0.38  0.81     0     0     0     0     6

## Float Columns
Column name: Non-nan   mean    std    min    25%    50%    75%    max
Age        :     714  29.70  14.53   0.42  20.12  28.00  38.00  80.00
Fare       :     891  32.20  49.69   0.00   7.91  14.45  31.00  512.33

## Other Columns
Column name: Non-nan   unique   top (count)
Name       :     891      891   Goldschmidt, Mr. George B (1)
Sex        :     891        2   male (577)
Ticket     :     891      681   347082 (7)
Cabin      :     204      148   C23 C25 C27 (4)
Embarked   :     889        4   S (644)