This project is a submission for the Udacity Data Science course. The main goal was to prepare the input messages and categoris files, perform the cleaning, find and build the best performing model, and output the results using Flask. The main output is the website with the model implemented under the hood, which is classifing the message into one or more out of 26 disaster categories.
There are 4 main catalogs in the repository: app, models, process_data, and tests.
The app folder contains all the files related to setting up the flask app, so the underlying model can be run on the web browser.
The models folder contains the train_classifier.py file which holds all the steps required for training the classification model.
The process_data folder contains all the steps required for the dataframe cleaning and returns merged datasets, ready for classification training.
The tests model holds all the unit tests.
-
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, run this line in the
disaster_responsefolder:
python process_data/process_data.py process_data/disaster_messages.csv process_data/disaster_categories.csv ./DisasterResponse - To run ML pipeline that trains classifier and saves run this line in the
modelsfolder:
python train_classifier.py ../DisasterResponse.db ../classifier - To run the tests, run this line in the main folder:
pytest
- To run ETL pipeline that cleans data and stores in database, run this line in the
-
Run the following command in the main directory to run your web app.
python app/run.py -
Go to http://0.0.0.0:3001/