| NAME | NUID |
|---|---|
| Vyshnavi Pendru | 002919813 |
| Moksha Ajaykumar Doshi | 002922797 |
Project Proposal link - Google Codelabs - https://codelabs-preview.appspot.com/?file_id=1qNgCSdMase7BZ0hIgzXX3MoG30t2BKT-CcIxz5PYo1c#8
Project Documentation link- Google Codelabs - https://codelabs-preview.appspot.com/?file_id=1OLQHLza5rEUGrf1w3VChJf5XYeAgSyagsfMOCwpWkmI#0
Project Demo link- https://drive.google.com/drive/folders/1R32qu0MoUBC8avR23ProxYCg6grsVlOo
Project link which is deployed on aws ec2 instance-https://test.nedamg7245fall2022.com/
To Predict whether a given transaction is fraud or not, we have created an streamlit application with backend using fastapi and authenticating with OAuth2 Authentication. Then dockerized both frontend and backend by deploying on AWS EC2 instance .
Steps to Run the Application
- Install Docker and Docker Compose
- git clone https://github.com/vyshnavi-pvr/damg-project
- cd damg-project
- add credentials in api/.aws and streamlit/.aws
- sudo docker-compose up -d --build
- The streamlit application runs on 8501 port and fastpi run on 8001 port
Storing data,model in AWS S3 bucket and retrieving it, Docker , Authentication and Prediction in FastAPI, Authentication in Streamlit, Deploying on AWS - Vyshnavi Pendru
Validation of data using Great Expectations, Data Science, EDA, Modeling & Analysis, Model-as-a-service in FastAPI, Analysis & Model in Streamlit - Moksha Doshi
docker-compose.yml
+---api
Dockerfile
model-as-a-service.py
model.pkl
requirements.txt
run.sh
+---.aws
config
credentials
+---datascience
create_model.py
credit_card_eda_models.ipynb
datascience.ipynb
model.pkl
pickle_model.pkl
+---Great_expectaion_on_data
ge_card_trans.html
+---streamlit
Dockerfile
login.py
Main_Page.py
requirements.txt
+---.aws
config
credentials
+---pages
Data_Statistics.py
Manually_Checking_Predictions.py
Try_Different_Models.py
+---util
get_data_s3.py