Disaster Response Message Classification

Disaster Response Pipeline Project

A Flask app that classifies a disaster response message into 36 categories

Instructions:

First, download the data set from : https://drive.google.com/drive/folders/1HSbVyyXSXWuEHHogFxyFYx6SxngO6qmE?usp=sharing

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/

ETL Pipeline

• Loading data from csv files
• Prepare Clean data, re arrange categories.
• Save data into sqlite data base

ML Pipeline

• Load data from database
• Build model with multioutput multiclassification estimator (MultiOutputClassifier and KNeighborsClassifier), and use Gridsearch cv to get the best parameter for the model
• Training model
• Evaluating model - calculate the recall score, precision score and F1 score of the model for each categories
• Saving model - save trained model into a pickle file

#Libraries used:

• pandas
• numpy
• scipy
• SQLAlchemy
• nltk
• sklearn
• plotly

Motivation for the project

This project was engineered by udacity , This aim of this project is classify a disaster response message into different categories in other to recognise which response unit to be alerted and to response to the messages.

Files in the repository

• Disaster-response
• template
  • go.html - an htlml file with jinja code to display the classification of the message
  • master.html - an html file with base codes that extends its templates
• run.py - a script that loads the home page and routes to classify messages with the classifier
• data
  • process_data.py - a script that cleans data and save data in db,
  • DisasterResponse.db - a result from the process_data.py
  • disaster_messages.csv - data to be processed
  • disaster_categories.csv - data to be processed
• models
  • train_classifier.py - read in data, tokenize, build a model and save the model as a pickle file
• notebooks
  • ETL Pipeline Preparation.ipynb - step by step on how to process , read and save to db
  • ML Pipeline Preparatio.ipynb - step by step on how to build a model and improve it with GridSearchCv