The problem CRASHCAST (Customer Behavior Analytics) solves

Our service will help startups and small businesses to take better care of their customers and provide priceless customer support.
Starting a start-up is a drag especially if have a big mission and a vision at hand. Handling small nuisances like customer service, customer churn prediction and customer behaviour prediction takes the attention away from improving the service or product that the start-up is offering.
Outsourcing these tasks also comes at a cost of increased expenses which in turn diverts the company’s focus off of the product or service. Automating these tasks provides tremendous benefit to the company or start-up. Our service will consider all the important features in a dataset like CustomerID, country etc and provide deep analysis and churn prediction using a Deep learning neural network and Support vector machine algorithm which provides the company, valuable information about the customers to make customer service and analysis faster and more efficient.
Our Project involves various machine learning models which will helps the to overcome this problem very easily and cost effectively.
After a positive customer experience, more than 85 percent of customers purchased more. After a negative experience, more than 70 percent purchased less. So getting this wrong can prove a costly exercise.
Rather than rely on assumption, a business needs know exactly what makes the customer feel like they are receiving superior service.

From our project it will be very easy for the startups or small bussiness to focus on the specific groupof customers which plays a main role in their revenue generation.
This helps them to cut their marketing budget and focus on product and service improvement.

Challenges we ran into

Challenges we ran into
Integrating the ML project into the web application causes a lot of difficulties.
Developing a Proc file for Heroku Deloyement (To over come this situation we have watch a lot of youtube videos)
Integrating Plots into streamlit (To overcome this we had to debug every missing librarires using pip)
Integrating various version of different libraries in python is very difficult ( To overcome this we specified specfic version in requirement file)
Training our model was difficult so we had to use googles TPU