Disaster response pipeline

Project motivation

The current project is done in the framework of the Data Scientist Nanodegree at Udacity and includes two main goals:

• to develop a pipeline that handles disaster data provided by Figure Eight and builds multioutput classifier model predicting the category of a disaster message;
• to make an application that allows an user to predict a category for any new message.

Description of files

• ETL Pipeline Preparation notebook contains all the steps needed for data cleaning, handling and saving the final dataset into SQL database.
• ML Pipeline Preparation notebook contains all my experiments with different classifiers
• The folder data includes disaster messages (disaster_messages.scv), categories for these messages (disaster_categories.csv), the final database with cleaned and preprocessed dataset (DisasterResponse.db), and the script cleaning, merging, and saving data (process_data.py) (the script is made on the base of ETL Pipeline Preparation notebook)
• The folder models includes the customer transformer calculating some statistics for an input text (custom_transformers.py) and the script with the pipeline for the final model (train_classifier.py).
• The folder app includes templates and script for launching the application.

How to run the pipelines and the app

1. To run ETL pipeline that cleans data and stores in database: python data/process_data.py data/disaster_messages.csv data/disaster_categories.csv data/DisasterResponse.db

2. To run ML pipeline that trains classifier and saves: python models/train_classifier.py data/DisasterResponse.db models/classifier.pkl

3. Run the following command in the app's directory to run your web app: python run.py

4. Go to http://0.0.0.0:3001/

Libraries

To successfully reproduce my script on your machine, the following packages should be installed: pandas, nltk, numpy, sklearn, json, plotly, flask, sqlalchemy