A repository for project on github. After

git clone https://github.com/andyc1997/mda.git

please go to the corresponding directory and run:

pip install -r requirements.txt

As the app needs to be deployed on Heroku, it is re-posted on another repo:

https://github.com/andyc1997/mda_app

The jupyter notebooks in code are organized as follows:

• Scraping comments from the reddit post: reddit_post.ipynb
• Create temperature thresholds using reference distributions: temp_reference_dist.ipynb

• Clean meteorological data and mortality data for data visualization and mortality prediction model: data_processing.ipynb with helper functions wrapped in subset_helper.py

• Text mining for patterns in the reddit post: text_mining.ipynb
• Summary statistics and basic visualization for exploratory analysis: summary_statistics.ipynb
• Perform cross-validation to select the most appropriate models for mortality prediction: model_poisson.ipynb, model_gradient_boosting.ipynb, model_randomforest.ipynb

• Build and deploy an App for data visualization with differenet countries and mortality prediction model: mda_app
• The App is deployed at here https://mdaproject.herokuapp.com/
• If there is an application error, please refresh the webpage and try again (probably due to request time out)

Remark: As data are provided in data, the jupyter notebooks can be run in any order.

1. Monthly mortality statistics in Belgium
2. Meteorological variables for European countries
3. Reddit: 'Hell is coming': Western Europe braces for its hottest weather since a 2003 heat wave killed 15,000 people in France.