MLAAS: use streamlit for ML deployment

Using Streamlit, we deploy  exploratory data analysis (EDA) for different tabular dataset.

The following data are preloaded in the model:

• Iris
• PIMA
• Wine
• Helth insurance
• External

The External option, will give the opportunity to upload csv file or URL to fetch the tabular data on-the-fly. The are 4 different tasks in the app:

• Data review
• Visualization
• Modeling
• Prediction

In the data review, the option to preview the data is given. In addition, a report generated from panda is given to give some insight about the data. The visulization task, including the scatter plot, histogram, and group plots (box plot, correlation in seaborn, bar plot). The modeling will learn one of the following machine learning model to solve a classification task:

• Logistic Regression
• Random Forest
• Support Vector
• Naive Bayes
• Decision Tree
• K Nearest Neighbour
• Linear discriminant
• Latest learned model

The latest learn model will remember the session information after applying different configuration in training. The configuration includes sampling method and the train-test ratio. An ROC plot will also be shown for the trained model. Finally the prediction, will provide the class for of the test cases declared by the user given different models.

The code is available in Github link bellow:

Restricted content! log in or register for free.

Here is the online demo version (wait to be loaded):

In case of issue, use this link to view the demo version:

Restricted content! log in or register for free.