Skip to content

Time series analysis showing trend, seasonality, and periodicity decomposition; and forecasting using Facebook Prophet. The analysis makes extensive use of indexing data tools and of the Pandas and Datetime libraries.

License

Notifications You must be signed in to change notification settings

paocarvajal1912/Forecasting_Net_Prophet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Forecasting_Net_Prophet

This notebook shows time series analysis, emphasizing the use of indexing data on dates, and the forecasting of sales and Google searches using Facebook Prophet. The analysis shows how to decompose time series in trends, seasonality, and periodicity.

The time series tools are applied to marketing analysis of MercadoLibre, which is a lider providing online shopping in Latin America.

Technologies

The analysis is done in Google Colab at https://colab.research.google.com. The main technologies used are: Pandas, Holoviews, Facebook Prophet, Hvplot, Datetime, Numpy and Matplotlib inline.

Instalation Guide

If you don't have this tools, you need to install them:

  • !pip install pystan
  • !pip install fbprophet
  • !pip install hvplot
  • !pip install holoviews

The installation appear in the first cewll of the notebook.

Usage

The main file is the forecasting_net_prophet.ipynb Jupyter Notebook. You should open it in Google Colab and run it completely in order to see the graphs.

For the upload of the files, the files should be selected during the run. There are three files to upload, and the cells to do it are the ones with this code:

from google.colab import files

uploaded = files.upload()

The files are in the Resources folder, and needs to be selected on the spot.

Some charts you will see in Google colab once you run the notebook are as follows.

Seasonality Analysis of the Mercado Libre Google Search Data

Logo

Logo

Stock Close Price and Google Search Trends Subplots

Logo

Logo

Logo

Logo

Logo

Logo

Logo

Contributors

This project was coded by Paola Carvajal Almeida, paola.antonieta@gmail.com.

Contact email: paola.antonieta@gmail.com LinkedIn profile: https://www.linkedin.com/in/paolacarvajal/

License

This project uses a MIT license. This license allows you to use the licensed material at your discretion, as long as the original copyright and license are included in your work files. This license does not contain a patent grant, and liberate the authors of any liability from the use of this code.

About

Time series analysis showing trend, seasonality, and periodicity decomposition; and forecasting using Facebook Prophet. The analysis makes extensive use of indexing data tools and of the Pandas and Datetime libraries.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published