From Open Data to App Store – Testing Helsinki’s Traffic API with Codex
The Helsinki Traffic Accident Map began as two experiments running side by side. On one level, I wanted to see what Helsinki’s open API datasets actually allow you to build in practice – not just pulling raw figures, but constructing a full pipeline from API call to cleaning, filtering and interactive visualisation. The city makes its traffic accident data available openly and at no cost, which makes it an ideal civic testbed. At the same time, this was a practical trial of Codex in a more agentic mode. In earlier projects I worked through iterative coding dialogues and handled the surrounding implementation steps myself, see Threat assessment monitoring tool and Pulling YouTube scripts . Here, the intention was to see how far the system could carry the process from idea to working prototype.
We started with a local proof of concept on my machine and iterated repeatedly, especially around UI clarity, filtering logic, naming consistency and accessibility. Once the beta felt stable, we pushed it to GitHub and only then did it seem worth writing about. The map is therefore more than a visualisation of accident data; it is a feasibility test. Having established that open municipal data and an agent-driven workflow can produce something coherent and inspectable, the next step is more ambitious: to move beyond a browser-based exploration tool and develop a downloadable civic application distributed via an app store. That transition brings new demands – reliability, maintenance, compliance and long-term support – but it also opens the possibility of creating something Helsinkians might actually use in everyday life.
Building with these systems inevitably raises questions about the broader ecosystem behind them. Recent developments around OpenAI’s licensing of its models, particularly in relation to military use, have given me pause. My initial reaction is one of discomfort. At the same time, I want to understand the details and implications more fully before arriving at a considered position.
What open data?
The city of Helsinki maintains an open data portal where the city publishes datasets for public use free of charge. These datasets, accessible via APIs, cover areas such as transport, environment, urban planning and safety, enabling developers, researchers and citizens to build applications and analyses on top of municipal data.
Data source
This project uses traffic accident data published by the City of Helsinki via its open data portal as described above. The map and visualisations presented here are an independent experiment and are not affiliated with or endorsed by the City of Helsinki.
While the data originates from official municipal datasets, the interpretation, filtering and presentation are my own. No guarantee is made as to completeness or accuracy, and the tool should not be relied upon for official, legal or safety-critical decisions.