Disaster Response Pipeline Project

Table of Contents

1. Description
2. Getting Started
3. Project Organisation

Description

This Project is part of Data Science Nanodegree Program by Udacity in collaboration with Figure Eight. The objective of the project was to build a tool that would classify disaster messages into categories - e.g. "water", "fire" or "food", to aid responses. The dataset used to train the model contains pre-labelled request messages sent during a natural disaster.

The project is divided in the following sections:

1. ETL Pipeline:

• Loads the messages and categories dataset
• Merges the two datasets
• Cleans the data
• Stores in a SQLite database

2. ML Pipeline - builds a model able to classify text message into categories

• Loads data from SQLite Database
• Splits the data into training and test setse
• Builds a text processing and ML pipeline
• Trains and tunes a model using GridSearchCV
• Outputs result on the test set
• Saves the final model as a pickle file

3. Flask Web App

• Allows user to send request messages to model and receive response in real time

Getting started:

1. Running the ETL Pipeline:

   • From the project directory run python data/process_data.py data/disaster_messages.csv data/disaster_categories.csv data/DisasterResponse.db

2. Running the ML Pipeline:

   • From the project directory run python models/train_classifier.py data/DisasterResponse.db models/classifier.pkl

3. Starting the Flask Web App:

• From the project directory run python /app/run.py
• Go to http://127.0.0.1:3000/

Project organisation:

Project Organization

├── app
│   ├── run.py------------------------# python file used to run the flask app
│   └── templates
│       ├── go.html-------------------# classification result page on app
│       └── master.html---------------# main page on app
├── data
│   ├── DisasterResponse.db-----------# database used for saving and loading data
│   ├── disaster_categories.csv-------# input data - message labels  
│   ├── disaster_messages.csv---------# input data - messages
│   └── process_data.py---------------# python file used to load and clean data
├── screenshots-----------------------# screenshots used in README
├── models
│   └── train_classifier.py-----------# builds and trains model used in web app