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.
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
.
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.
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.
This project was coded by Paola Carvajal Almeida, [email protected].
Contact email: [email protected] LinkedIn profile: https://www.linkedin.com/in/paolacarvajal/
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.