datamallet 0.22.0

Helper tools for the data science workflow

datamallet

Datamallet is a collection of helpful functions and modules built by Data scientists for Data scientists, to help expedite the data science workflow.
From a technical standpoint, datamallet is built on top of the following libraries:

1. Scikit-learn (for creating the transformer classes).
2. Plotly (for the automatic visualization function).
3. Pandas (for creating scikit-learn compatible transformers, and for creating utility functions for wrangling data).
4. Numpy
5. Scipy.

The goal of this project is to help Data scientists become more efficient in their roles by providing commonly used functionality that have been battle tested and have been contributed by other Data scientists.

Installation

datamallet is available on pip and can be installed using the command below:

pip install datamallet

Tests

from the main directory, you can run the tests by simply running the pytest command.

pytest

Quick start

Automatic Data Visualization

from datamallet.visualization import AutoPlot
import plotly.express as px
from datamallet.visualization import AutoPlot

tips = px.data.tips()

autoplot = AutoPlot(df=tips,include_scatter=True,include_pie=True,include_box=True, include_sunburst=True, include_violin=True, include_treemap=True, include_histogram=True, include_correlation=True, create_html=True, filename='autoplot')

list_of_charts = autoplot.show()

An html file with filename autoplot.html would be created (check the current directory, a sample is also found in this repo), the show method also creates a list of the plotly graph objects, so you have the option of not creating the chart but using the list of graph objects to display the charts.

for chart in list_of_charts:
chart.show()

Modules

datamallet currently has the following modules

1. Visualization module which contains helper functions for automatic visualization and for creating different types of charts such as:
   -Scatter plots.
   -Correlation plots.
   -Histogram.
   -Box plots.
   -Violin plots.
   -Treemaps.
   -Sun burst Charts.
   -Pie Charts.
   -Density Contour Charts.
   -Density Heatmap.
   

All these charts can be created automatrically using the Autoplot class in the visualization module, they can also be created using individual functions in the plot module.

2. Tabular module contains scikit-learn compatible transformers for data manipulation for tabular data, (data which can be found in a table (pandas dataframe) either pure tabular or timeseries). The classes found in the tabular module can be used in a scikit-learn pipeline.
   The Tabular module contains the following submodules:
   -features which contains scikit-learn compatible transformer classes for creating new features (more classes are welcome).
   -timeseries which contains transformers for manipulating time series data.
   -utils which contains helper functions for data wrangling and carrying out checks.
   

• preprocess which contains transformers for preprocessing data.