In this job, I collaborated with Ahmed BESBES

You may read about it here and here.

In this post, we'll go through the necessary steps to build and deploy a machine learning application. This starts from data collection to deployment; and the journey, you'll see, is exciting and fun. 😀

Before we begin, let's have a look at the app we'll build:

As you see, this web app allows a user to evaluate random brands by writing reviews. While writing, the user will see the sentiment score of his input updating in real-time, alongside a proposed 1 to 5 rating.

The user can then change the rating in case the suggested one does not reflect his views, and submit.

You can think of this as a crowd sourcing app of brand reviews, with a sentiment analysis model that suggests ratings that the user can tweak and adapt afterwards.

To build this application, we'll follow these steps:

• Collecting and scraping customer reviews data using Selenium and Scrapy
• Training a deep learning sentiment classifier on this data using PyTorch
• Building an interactive web app using Dash
• Setting a REST API and a Postgres database
• Dockerizing the app using Docker Compose
• Deploying to AWS

To run this project locally using Docker Compose run:

docker-compose build
docker-compose up

You can then access the dash app at http://localhost:8050

If you want to contribute to this project and run each service independently:

In order to launch the API, you will first need to run a local postgres db using Docker:

docker run --name postgres -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=password -e POSTGRES_DB=postgres -p 5432:5432 -d postgres

Then you'll have to type the following commands:

cd src/api/
python app.py

In order to run the dash server to visualize the output:

cd src/dash/
python app.py

Feel free to contribute! Report any bugs in the issue section.

Here are the few things we noticed, and wanted to add.

• Add server-side pagination for Admin Page and GET /api/reviews route.
• Protect admin page with authentication.
• Either use Kubernetes or Amazon ECS to deploy the app on a cluster of containers, instead of on one single EC2 instance.
• Use continuous deployment with Travis CI
• Use a managed service such as RDD for the database

MIT