Rill Workshop Instructions

📊 Slide Deck: View the workshop presentation

Creating your git commits project

1. Install Rill:

   curl https://rill.sh | sh

   Start Rill:

   rill start rill-git-analytics

2. Rill opens in your browser. Click on Start with an empty project

3. Enable AI Features in Rill Developer (Beta)

   Add the following to rill.yaml:

   features: 
       developerChat: true
       generateCanvas: true

4. Add -> Data (Top left) → Select Google Cloud Storage → Skip credentials part (It's a public source)

   Location:

   gs://rilldata-public/db_repos/commits.json.gz
   

5. Wait for data to load (might take 1-2 mins depending on network speed)

6. Generate dashboard using AI

Get familiar with core concepts

Check and review project files one by one:

1. Project Metadata (rill.yaml) - project level metadata
2. Sources (sources/commits.yaml) - source definition for reading data from GCS to DuckDB
3. Metrics View (metrics/commits_metrics.yaml) - Dimensions, measures, security policies
4. Explore (dashboards/commits_metrics_explore.yaml) - Exploratory dashboard for commits
5. Canvas (dashboards/commits_metrics_canvas.yaml) - Report style canvas dashboard

Quick Edits using AI as copilot

Use prompt below -

Add a new dimension `commit_uri` that points to commits on GitHub
Add a new measure for the number of committers as count distinct authors

1. Notice the tool calls being made
2. Explore these new dimensions and measures
3. Find 1 insight and share it
4. Popular projects, Top contributors
5. Compare activity across projects over time

Editing code outside of Rill

1. Open project in your fav IDE/CLI
2. Add a new file models/file_changes.sql
3. Copy contents from file_changes.sql
4. Create a new Metrics View in the same way
5. file_changes_metrics.yaml
6. Que: What is the lag between adding the file and it reflecting in Rill
7. BI-as-Code allows any AI-enabled IDE for making changes

Deploy your project to Rill cloud

1. Click on Deploy button top right
2. Need signup if have not previously used Rill

Try conversational BI in rill cloud

1. Who are top 10 committers for clickhouse repository in last 3 months
2. Which commit deleted the most amount of lines?
3. Can you spot this commit as anomaly on dashboards?
4. Which repository has the highest number of commits in last 1 year?
5. Show me number of commits over time

Further tune AI behavior by adding more business context

1. Add business rules additional ai_instructions at project or metrics view level
2. e.g. My financial year starts on April 1
3. Refer to AI Configuration documentation

Fun things to try:

1. Add ai instruction for AI to respond in hindi or any other language of your choice. (Multi-lingual support)
2. Ask AI to add respond more funnily
3. Ask AI to praise you in every response
4. Add thanking notes
5. Share the most fun/unexpected response you got :)

Further exploration

Your own git commits data

• Use script in scripts/extract_commits.py to extract commits from one or more repositories

Connect using MCP server

• MCP Server Documentation
• Top 10 Tips for Using Claude Desktop with Rill Data

Advanced Security policies

• Security Policies Documentation

Embedding Rill in your applications

• Embedding Documentation

More demo projects

1. Check out public Rill Demo Projects
2. Example projects on GitHub

---

Get Involved

1. Like & Star the Rill project on GitHub if you found it useful
2. For any questions reach on Discord or report issues
3. If you are building something similar, let's connect, collaborate and share ideas