wayneong95 / disaster-response-pipeline-project

A submission for the Udacity Data Scientist Nanodegree program

Table of Contents

1. Installation
2. Project Summary
3. File Descriptions
4. Instructions

Installation

There should be no necessary libraries to run the code here beyond the Anaconda distribution of Python. The code should run with no issues using Python versions 3.*.

Project Summary

A web app that classifies disaster messages into categories so that the right help will reach the people in need. An emergency worker can input a disaster message and get classification results in several categories. The web app also displays visualizations of the training data used.

File Descriptions

There are three folders - app, data and models.

The app folder contains run.py which runs the web app and a templates folder that contains the html files - go.html and master.html.

The data folder contains the raw csv files - disaster_categories.csv and disaster_messages.csv, and the process_data.py file, which loads, cleans and save the cleaned data to a sqlite database.

The models folder contains train_classifier.py, which builds, trains, evaluate and save the trained model as a pickle file.

Instructions

First, run the ETL pipeline to clean and store the data in the sqlite database via the command "python data/process_data.py data/disaster_messages.csv data/disaster_categories.csv data/DisasterResponse.db".

Next, run the ML pipeline that trains the classifier and saves it as a pickle file via the command "python models/train_classifier.py data/DisasterResponse.db models/classifier.pkl".

Run the web app via the command "python run.py".