Primarily, this project is deisgined to use various ML algorithms to predict the weekly sales profit for a given input week based upon the training dataset. The output of this prediction, comes with the revenue_value along with its accuracy metrics and a sactter plot to depict its general accuracy!
We can also use the add_product or add_dataset endpoints to either add slowly accumulating data which at the end of each week is aggregated to calculate each week's sales_profit value, or we can just add a single dataset of weekly sales_profit directly!
Backend api server and database(default sqlite)
• To start the backend server:
python manage.py runserver
To aggregate the weekly sales_data
To use aggregation functionality
• First install "Erlang"(required for RabbitMQ) and then, "RabbitMQ-Server"(A task queue)
• Basic requirements- Open 2 new terminals with each running the following commands:
celery -A back_end worker -l info --pool=solo
celery -A core beat -l INFO --scheduler django_celery_beat.schedulers:DatabaseScheduler
To predict the sales value for a particular date using various ML algorithms
• Used "sci-kit learn"(sklearn) and "xgboost" packages
Frontend framework used along with Django
• To start the frontend server:
npm start