https://github.com/drearondov/nlp-newspapersanalysis

End-to-end NLP Project of news headlines for the main newspapers of Perú
https://github.com/drearondov/nlp-newspapersanalysis

eda kedro nlp python sentiment-analysis topic-modeling

Last synced: about 2 months ago
JSON representation

End-to-end NLP Project of news headlines for the main newspapers of Perú

• Host: GitHub
• URL: https://github.com/drearondov/nlp-newspapersanalysis
• Owner: drearondov
• License: bsd-3-clause
• Created: 2021-03-07T23:24:58.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2024-01-03T23:15:18.000Z (about 1 year ago)
• Last Synced: 2024-05-12T00:46:52.599Z (8 months ago)
• Topics: eda, kedro, nlp, python, sentiment-analysis, topic-modeling
• Language: Jupyter Notebook
• Homepage: https://drearondov.github.io/NLP-newspapersAnalysis/
• Size: 9.2 MB
• Stars: 2
• Watchers: 0
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# NLP - Peruvian Newspapers Analysis

![Python Version](https://img.shields.io/badge/python-%3E=3.9-blue?style=for-the-badge&logo=python&logoColor=white)

[![Powered by Kedro](https://img.shields.io/badge/powered_by-kedro-ffc900?logo=kedro&style=for-the-badge)](https://kedro.org)

![Code style badge](https://img.shields.io/badge/style-black-black?style=for-the-badge)

[![GitHub deployments](https://img.shields.io/github/deployments/drearondov/NLP-newspapersAnalysis/github-pages?style=for-the-badge&label=Documentation)](https://drearondov.github.io/NLP-newspapersAnalysis/)

A full Data Science pipeline project about the analysis of newspapers headlines

of Perú.

## Objective

To show how the narrative changes over a period of time in local media, from both

independent and main news outlets in Perú.

### Specific Objectives

- Show asociations between different words over a period of time to see how the

  narrative changes around ceirtain topics

- Show how the media can control the narrative, by looking into the different

  reactions people have to the tweets

## Project overview

The project consists of 6 pipelines that go from retrieving the data to the

data structures that can be used for Data Analysis and Visualisation. Below is

a picture of the pipelines flow.

![pipeline flow for the project](docs/_static/nlp-newspaper-pipeline.svg)

> You can also check the sister project [nlp-newspapersDashboard](https://github.com/drearondov/nlp-newspapersDashboard)

> where I buil a dashboard with the data comming from this project.

### Pipelines

- Data Retrieval

- Cleaning and Preprocessing

- Feature Engineering

- EDA

- Sentiment & Emotion Analysis

- Topic Modeling

Full documentation about the project and the data warehouse can be found in

the [documentation](https://drearondov.github.io/NLP-newspapersAnalysis/). And

if you want to learn more abouth the building process you can read the

accompaning blog post series on [Nou de Data](noudedata.com).

## Code and Resources used

- **Python version:** `3.11.5`

- **Packages:** Kedro, Pandas, Numpy, Plotly, Requests, Gensim,

  Textblob, PySentimiento