No-Code Platform Development
This platform is designed to help users navigate through the end-to-end predictive analytics using AI/ML tools, without the need for programming expertise. It features a point-and-click interface with user-guided feedback, and is specifically tailored for public health researchers.
The No-code platform has 3 main components:
Data Processing
This component implements basic techniques to process and visualize data, including: Data upload, Data validation, Descriptive statistics, Feature engineering and Data partitioning.
Modelling
Provides users with a flexible and guided stepwise workflow of the modelling process, i.e. Model selection, Model training and tuning, Model evaluation and Model external validation and comparison.
Rollout
Integrates a trained model into a production and monitoring pipeline of: Deployment, Serving, Prediction and Inference.
The No-Code Platform Development aims to:
1. Build Capacity
- Deploy the existing I-DAIR Codex for Pathfinders.
- Conduct platform installation and training sessions to facilitate platform adoption.
2. Develop the Architecture for APHRC Codex Platform
- Build a new, flexible version in R, with some components in python for specific ML/AI tasks, while incorporating Pathfinder feedback.
- Integrate the platform with the OHDSI pipeline, utilizing the OMOP Common Data Model (CDM).
3. Enhance Analytical & Visualization Capabilities
- Incorporate advanced tools for analysing longitudinal data and AI/ML models.
- Incorporate robust data management and visualization modules.
