Reproducible pipeline from Twitter API using DVC

In this project I built a pipeline using DVC from my previously created notebook, called the Twitter API.
Due to the size of my notebook, I only put the most important parts of my work into the pipeline.
This parts are:

• Creation of dataset (I reduced the size of the dataset due to time reasons)
• Creation of a NetworkX graph
• Generation of the image of the graph

The pipeline graph is the following:

  +-------+
  | fetch |
  +-------+
       *
       *
       *
  +-------+
  | graph |
  +-------+
       *
       *
       *
+------------+
| egonetwork |
+------------+

Setup

Download

To download the project, proceed with cloning.

git clone https://github.com/antoniod20/dvc-twitter.git

Configuration

The project was carried out with Python 3.6.9. It is therefore advisable to have a version of Python at least higher than version 3 installed.
To install all the libraries needed to run the project, it is necessary to run this command line:

pip install -r src/requirements.txt

Run

To launch the pipeline, the following steps must be run:

• First command

  cd .\dvc-twitter\dvc-twitter-api\

• Second command

  dvc repro

Resources & Libraries

• Tweepy - Twitter API
• NetworkX - Useful to handle the study of graphs and networks
• Pandas - Useful to handle the CSV file
• Matplotlib - Provides functions for embedding plots into applications

Author

• Antonio Donvito