zeitpy 0.1.2

A package for time series analysis — from EDA to forecasting and performance assessment.

ZeitPy

ZeitPy is a package for time series analysis. Its main purpose is to abstract some vital operations to analyse time series data. At its core is the Zeit class, which provides attributes and methods for initializing a (Pandas) time series, performing Exploratory Data Analysis (Augmented Dickey-Fuller test, visualize periodograms, seasonal, and lag plots, etc.), forecasting and performance assessment.

---

Class Overview ℹ️

Zeit

Class Initialization

In [1]: import zeitpy as zp
...: zo = zp.Zeit('sales_luanda.csv', date_format = '%Y-%m-%d', date_col = 'date', data_col = 'sales')

• dataset: the DataFrame, Series or csv file path containing the time series data.
• date_format: the format of the "date_col" instances to be converted into datetime.
• date_col: the column containing the time observations.
• data_col: the column containing the values (in case of csv files or DataFrames).

Attributes and Methods

• data

  • this attribute retrieves the time series data wrapped by the Zeit object.

• seasonal_decomposition(model: str = 'additive', period: int = 12, **plot_args) -> None

  • A method for plotting the seasonal decomposition of the time series using moving averages.
  • Parameters:
    • model: the type of seasonal decomposition
    • period: period of the series (12 for monthly data, 1 for annual, etc.)

---

Usage ⚙️📦

How to get the package?

Installation via PyPI

pip install zeitpy

Installation of source distribution (.tar.gz)

pip install path/to/zeitpy-0.1.2.tar.gz

Installation of pre-built distribution (.whl)

pip install path/to/zeitpy-0.1.2-py3-none-any.whl

Cloning the package repository

git clone https://github.com/domingosdeeulariadumba/ZeitPy.git

Importing the package

In [1]: import zeitpy as zp

Example: Injecting a dataset and accessing the first five records of the time series

In [2]: zo = zp.Zeit('sales_luanda.csv', date_format = '%Y-%m-%d', date_col = 'date', data_col = 'sales')
...: zo.data.head()
Out[2]: 
2024-09-07     86662
2024-09-08    449329
2024-09-09     64041
2024-09-10    420328
2024-09-11    351528
Freq: D, Name: sales, dtype: int32

📑 You can view all operations provided by the Zeit class implemented here.

---

License ⚖️

This project is licensed under the MIT LICENSE.

---

Issues and Contributions 🧱

Feel free to submit any issue you may find in this package or recommend additional features not already listed here. If you find this useful, please fork the repository, create a feature branch, and submit a pull request.

---

Connect with me 🌐

/domingosdeeulariadumba

/domingosdeeulariadumba

/domingosdeeulariadumba