Project for creating an Natural Language Processing (NLP) pipeline for recognizing a disaster response.

1. Project Motivation
2. Installations
3. File Descriptions
4. How To Interact With the Project

One of the most important challenges in an emergency is the speed at which we respond to a disaster. The quicker the response, the quicker people are able to help in the case of a disaster. Thanks to the growth of social media platforms and the exponential growth of textual data, we are able to identify the presence of an emergency within seconds.

Using natural language processing, we can categorize text to identify the type of help that is required. Although this is an immense undertaking, we can start by categorizing disaster responses into different subcategories.

For this data science project, the following libraries are required:

• JSON
• NLTK
• Flask
• Regex
• Numpy
• Plotly
• Pandas
• Pickle
• Sklearn
• SQLAlchemy

Use "pip install" to download the libraries.

The project consists of three main folders - app, data and model.

1. The app folder contains the HTML files, and the python file for deploying the Flask app.

2. The data folder contains the DisasterResponse database as well as the accompanying CSV files that contain the raw data. The process_data.py cleans the raw CSV files, and creates a database for the disaster response data.

3. The model folder contains the code that will extract information from the database, and fit the data onto a machine learning pipeline and output an NLP model. This model will classify text as different disaster responses.

If you would like to create the NLP model, use "git clone" to download the project. Then run the project in the sequence of process_data.py, train_classifier.py and finally the run.py file to deploy the application.

The application can be hosted on the local machine, or can be hosted on Platform-as-a-service such as Heroku.