NLP_pipeline_model

Udacity Project: Disaster Response Pipeline

In this project I will create a model which will classify disaster response messages using machine learning algorithms.

Process:

The following folders and processes are used:

Data

process_data.py: reads the data, cleans and uploads it to a SQL database. Basic usage is python process_data.py MESSAGES_DATA CATEGORIES_DATA NAME_FOR_DATABASE Datasets are disaster_categories.csv and disaster_messages.csv DisasterResponse.db: resulting database from transformed and cleaned data.

Models

train_classifier.py: has the code to load data, transform it using NLP processing, then run a machine learning model using GridSearchCV and train it. Basic usage is python train_classifier.py DATABASE_DIRECTORY SAVENAME_FOR_MODEL

App

run.py: Contains code for a Flask app deployment and the user interface used to predict results and display them.

ETL Pipeline Preparation and ML Pipeline Preparation are two jupyter notebook study files that were used for beta version of the code.