Chapter 5: Quickstart Tutorials
This chapter provides hands-on tutorials that walk you through the core capabilities of the MATIH Platform. Each tutorial is self-contained and can be completed in 15 to 30 minutes. By the end of this chapter, you will have experience with natural language querying, dashboard building, model training, data quality exploration, and platform administration.
Prerequisites
Before starting these tutorials, ensure you have:
- A running MATIH instance -- either local (see Local Development Setup) or cloud-deployed (see Cloud Deployment)
- First-time configuration completed -- an admin user, a tenant, and at least one data source (see First-Time Configuration)
- A web browser -- Chrome, Firefox, or Edge (latest versions)
Tutorial Overview
| Tutorial | Workbench | Time | What You Will Learn |
|---|---|---|---|
| Your First Natural Language Query | Agentic Workbench | 15 min | Ask questions in plain English and get SQL results |
| Building a Dashboard | BI Workbench | 20 min | Create visualizations and arrange them in a dashboard |
| Training a Model | ML Workbench | 25 min | Train, evaluate, and register a machine learning model |
| Data Quality Exploration | Data Workbench | 20 min | Profile data, detect anomalies, and set quality rules |
| Platform Administration | Control Plane UI | 20 min | Manage tenants, users, roles, and monitor the platform |
Sample Data
The tutorials use a sample retail analytics dataset that is included with the platform. The dataset contains:
| Table | Rows | Description |
|---|---|---|
orders | 50,000 | Customer orders with timestamps, amounts, and status |
customers | 5,000 | Customer profiles with demographics and segments |
products | 500 | Product catalog with categories and pricing |
inventory | 2,000 | Current inventory levels by warehouse |
returns | 3,500 | Product returns with reasons and resolution |
This dataset is automatically loaded when you create a tenant with the sample data option enabled, or you can load it manually through the Data Workbench.
Tutorial Flow
The tutorials are designed to be completed in order, with each one building on concepts from the previous:
[1. Natural Language Query] -- Learn conversational analytics
|
v
[2. Building a Dashboard] -- Visualize query results
|
v
[3. Training a Model] -- Predict outcomes from data
|
v
[4. Data Quality] -- Ensure data reliability
|
v
[5. Platform Admin] -- Manage the platformHowever, each tutorial can also be completed independently if you are only interested in a specific capability.
Accessing the Workbenches
Local Development
| Workbench | URL |
|---|---|
| Agentic Workbench | http://localhost:3003 |
| BI Workbench | http://localhost:3000 |
| ML Workbench | http://localhost:3001 |
| Data Workbench | http://localhost:3002 |
| Control Plane UI | http://localhost:3004 |
Cloud Deployment
| Workbench | URL Pattern |
|---|---|
| Agentic Workbench | https://agentic.{tenant}.matih.ai |
| BI Workbench | https://bi.{tenant}.matih.ai |
| ML Workbench | https://ml.{tenant}.matih.ai |
| Data Workbench | https://data.{tenant}.matih.ai |
| Control Plane UI | https://admin.matih.ai |
Getting Help
If you encounter issues during the tutorials:
- Check the Verification page for troubleshooting guidance
- Review service logs in the relevant workbench's developer console
- Use the platform status script:
./scripts/tools/platform-status.sh
Begin with Your First Natural Language Query.