GoodNews

A news app which delivers news which can be filtered by sentiment, if you want good news, you can have it

Oxford Hack 2020

This is a project created for the Oxford Hack 2020 by Angus Stanton and J Keane Quigley.

News Sources

The news articles are obtained from newsapi.org. The results are filtered by country (United Kingdom and United States) and also divided into catagories of general, sports, technology, science, health, business and entertainment if requested.

Sentiment Analysis

The sentiment analysis is done through Google Cloud's language library for python. Each news article is given a sentiment score ranging from -1 to 1. The higher the score, the more positive the artice is according to the analyser.

The App

The app is build with React-Native and is compatable with Android and iOS.

Requirements

To run the code in the newsData directory, the following requirements must be met:

Packages

The following packages are needed:

flask
sqlalchemy
google-cloud-language

To install these packages, run the following in your terminal:

pip3 install <package>

replacing <package> with the package name.

Google Cloud Setup

To use the Google Cloud Services you need to follow these instructions to set up a Google Cloud Natural Lunguage API Project and credentials.

News API Key

To access the news articles from newsapi.org, you must first sign up and get an API key.
The key should be copied and pasted into the file newsData/fill.py at the variable api_key.
Demonstratively, replace the <PASTE YOUR API KEY HERE AS A STRING> with your key:

...
from model import Base, AllNews, GeneralNews, SportsNews, TechNews, ScienceNews, HealthNews, BusinessNews, EntertainmentNews
from news import news


api_key = <PASTE YOUR API KEY HERE AS A STRING>
url = "https://newsapi.org/v2/top-headlines?pageSize={page_size}&country={country}&category={category}&apiKey={api_key}"
...

Run

Firstly chage directory in a terminal window to the newsData directory.
To run the app, the database must be populated with the news data. To do this run the following in the terminal window:

python3 fill.py

This will take some time to finish due to the large number of news articles being fetched and analysed.
Once it is finished, run the following in the terminal:

python3 api.py

This will allow you to access the data in the database in JSON format through your browser. Just use the url shown in the terminal window; likely http://127.0.0.1:5000.

Finding the Correct Data

The articles can be divided into categories and filtered depending on the sentiment analysis scores.
To look at category-specific data type http://127.0.0.1:5000/news/CATEGORY into your browser, replacing CATEGORY with the desired one.
To filter out articles with a sentiment score below a certain value, use http://127.0.0.1:5000/news/CATEGORY/sentiment/MIN_SCORE, replacing CATEGORY as mentioned as well as MIN_SCORE.
Remember, MIN_SCORE is a value in the range [-1, 1].
The available categories are:

all
general
sports
technology
science
health
business
entertainment

abstanton / GoodNews