Dashboards: the Low Hanging Fruit for AI

Dashboards are one of the lowest hanging fruits for analysts using AI. Once you have the data, AI can visualise it almost instantly.

China Gas Dashboard

With an AI agent, I created a dashboard which takes public JODI data and turns it into a usable view of China’s gas imports and calculated demand, that answers the questions:

• How do China’s imports compare with previous gas years?
• Is the change coming from LNG or pipeline supply?
• How does calculated demand compare with recent history?

Normally, building this would involve downloading the data, learning the relevant JODI flow codes, cleaning the file, writing transformation logic, producing charts, and creating a webpage to display everything.

With the AI agent, the first working version took less than 2 hours.

You can explore it here: China gas explorer

• It opens as a guided explanation (for those interested in the gas fundamentals) and can switch into a compact dashboard for repeat visits.
• It refreshes with new data releases from JODI according to their update calendar, so feel free to visit the page again next month.

The dashboard was the easy part. The more interesting problem came when the data I wanted did not exist: how AI Agent used emissions data to model China’s gas demand by sector.