Disaster Response Pipeline Project

Summary

This project is build a model to predict the disaster situations from text.
Project template and dataset is provided by Udacity Data Science Nanodegree - Project 1

How I train the Model?

• CountVectorizer & TfidfTransformer: Text data is vectorized by TF-IDF
• RandomForestClassifier: Multiple decision trees to predict the multi-label categories
• GridSearchCV: Search for best parameters. But due to the limitation of computational performance, I could not try various hyperparameters.

Setup

• Python3 required

Install packages

pip install -r requirements.txt

Instructions:

1. Run the following commands in the project's root directory to set up your database and model.

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

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

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

Project Structure

data/

• read data from csv files
• data cleaning, drop duplicates
• save cleaned data to sqlite database(data/DisasterResponse.db)

models

• read data from database and tokenize
• train models
• save trained models as pickle to models/classifier.pkl

app/

• flask app for running demo in web browser
• get the models from models/classifier.pkl and use it