The Missing Plate

The story behind The Missing Plate

Two international students. Two vegetarians. One of the world's greatest food cities.

What could possibly go wrong?

When we first arrived in Singapore for university, we were away from home and figuring out a completely new country for the first time. During orientation, we were taken to experience something we had heard about long before arriving: Singapore's iconic hawker culture.

We were excited. Very excited.

Then came the slightly awkward part.

We walked from stall to stall, scanning menus and asking questions, only to realise there was almost nothing we could eat.

At one point, we even found tiny fish in what we thought was our vegetarian hawker meal. Safe to say, that did not help.

It sounds like a small problem. But when you are new to a country, slightly homesick, and surrounded by one of the world's most famous food cultures, not being able to take part makes you feel surprisingly left out.

Our alternatives were often restaurants and cafés that cost significantly more than a hawker meal. And that got us thinking:

If we were struggling to find affordable vegetarian food, how many other people were facing the same problem?

That question became The Missing Plate.

From a missing plate to a market opportunity

We did not want to build another app that simply tells vegetarians where food already exists.

We wanted to find where it should exist next.

The Missing Plate uses data to identify hawker centres with few listed vegetarian options and uncover places where there may be an underserved opportunity.

Our idea is simple: help hawker operators make better decisions about where vegetarian demand may exist and what action they could take to serve it.

We combine information such as nearby population, existing vegetarian options, stall capacity and vacancies to identify promising hawker centres.

For each opportunity, the goal is to provide a practical next step: an existing stall could introduce an affordable vegetarian dish, or an operator could consider a vegetarian concept when a stall becomes available.

In other words, we want to turn missing plates into marketplace opportunities.

How we are building it

Our solution brings together publicly available data from multiple sources:

• NEA and data.gov.sg for hawker centre locations and cooked-food stall counts
• SingStat for nearby resident population data
• Public vegetarian listings to understand where vegetarian options already exist

Within the Databricks Lakehouse, raw datasets enter our Bronze layer before being cleaned and matched in the Silver layer. AI Functions can help extract useful evidence from listing text, while MLflow can be used to evaluate that extraction.

The cleaned data then feeds into our Gold layer, where opportunity scores can be generated using Databricks SQL. The final vision is a simple operator-facing app that shows promising locations on a map and, importantly, suggests the next action an operator could consider.

Because identifying a gap is useful.

Knowing what to do about it is better.

The challenge

Ironically, the biggest challenge in building a project about missing vegetarian data was... missing vegetarian data.

We initially hoped to find one clean dataset showing vegetarian stalls across Singapore's hawker centres.

It did not exist in the form we needed.

Instead of abandoning the idea, we changed our approach. We found other useful public datasets and designed the solution around combining them. Hawker locations, stall information, population data and public vegetarian listings can each tell us one part of the story.

Together, they can help us see a gap that no single dataset shows on its own.

That challenge actually strengthened the idea. Real-world problems rarely arrive as perfect CSV files. We learned that sometimes the value is not in finding the perfect dataset, but in connecting imperfect ones intelligently.

What we learned

We started this project because we could not find lunch.

We ended up learning something much bigger.

A frustrating personal experience can point to a genuine market gap. Data can help turn that frustration into something useful, not just for people like us, but also for the hawkers trying to understand where new demand may exist.

Most importantly, inclusion does not always require reinventing an entire system.

Sometimes it can start with one more affordable dish on a menu. One better-informed operator. One missing plate finally being served.

What's next?

Our next step is to strengthen the vegetarian listing data, refine how we score opportunities, and validate whether the gaps identified by the model reflect real demand on the ground.

Over time, the same approach could potentially be adapted to other underserved dietary needs and communities.

Singapore's hawker culture is one of the things that makes the country special.

*We just want to make sure there is a plate for everyone. *

Built With

• data.gov.sg
• databricks
• lakeflow
• lakehouse
• singstat