Late one night, analysing BikeSpot 2023 data (almost 12 months ago), I found myself with four browsers open across two screens. One tab held a New South Wales crash portal; the others were pulling data for the City of Ballarat and the City of Bendigo. I was finishing BikeSpot reports that merged thousands of public BikeSpot submissions with official crash records.
Every portal had its own export quirks, its own field names, its own breakdown of ‘severity’. Minutes stretched into hours while I tried to reconcile the datasets so the story made sense for each council. I realised that if this is how long it takes me, someone who lives in civic data every day, then councils and advocates would be struggling too. That was the moment CrashDash stopped being an idea and became something that needed to be built.
In this post:
Why BikeSpot 2023 proved Australia needs a unified crash dataset
How messy spreadsheets evolved into the national CrashDash platform
What’s next: national coverage, Pro tier analytics, and human-centered AI
BikeSpot 2023 Project
What Needed to Exist
Local governments, transport planners, and community advocates face a simple question every day: where are the riskiest parts of our road network, and what is changing over time? Yet every state and territory publishes crash data differently. Some rely on static PDF summaries, others hide it behind search-driven portals designed for specialists. The format, access levels, and timing of releases shift state by state. The friction is so high that only people with the time and specialist training can make sense of the data.
From a civic tech perspective, that situation never sat right with me. Crash data is publicly funded and should be usable. It should not take hours to figure out what happened last quarter on a main road. It should not require a GIS license to see patterns that could prevent crashes from happening. The reality on the ground, though, was that the data sat in silos, and the tools that existed were either tucked away inside agencies or clunky enough to deter regular use after each update. I wanted CrashDash to feel like the opposite of that experience: a national crash map that should exist.
Many State portals - CrashDash
Building Toward a National View
CrowdLab evolved from my earlier work running CrowdSpot, a community engagement and mapping practice. Those projects showed a consistent theme: when you present complex civic information in a format people can actually use, they step in and help improve their community. With traffic safety, the stakes are even higher. Councils and residents are already motivated. They just need a clear, trusted picture.
The first prototypes were messy spreadsheets and hacked-together map tiles. Still, the first time I dropped Victoria and New South Wales into a single view, I knew I had to complete the picture. When I added Queensland and Western Australia a few weeks later, the tone of those conversations changed. This wasn’t just tidy data prep. It was the beginning of a national transport analysis platform. The breakthrough was realising that open data reforms alone would not solve the problem. Someone had to wrap those releases in a single experience people could access and depend on.
Design Principles That Guide Us
The first principle is the user experience. CrashDash had to be fast, reliable, and friendly enough that anyone could use it minutes after loading the page. Open data only fulfils its promise when people can actually work with it. If a council officer is standing in a meeting, they shouldn’t be waiting for maps to load or wrestling with obscure filters. So I obsessed over performance, the user experience, and plain-language filters.
The real proof came recently when providing a demo to a local council. Their use case, ‘walking and bike-riding crashes within 500 metres of a primary school, an hour either side of the bell’, was able to surface answers within a couple of minutes. That moment proved the product could unlock answers that once took days of specialist effort.
The second principle is coverage. The product had to offer a national lens. Anything less just rebuilds the silos we are trying to remove. Councils along state borders, consultants working across regions, and national advocates all need the same cohesive view. That is why each state and territory is on the roadmap, and why we keep pushing the ingestion pipeline to handle every dataset consistently.
The third principle is transparency. If I’m asking people to trust this tool, I have to show when each dataset was last updated, include definitions, and provide additional context. We’re building state-by-state explainers, legal protections, and clear documentation that anyone can read.
The fourth principle reaches into the future. I’m drawn to human-centered AI. I see CrashDash as a platform where machine learning assists rather than overwhelms. That could mean smarter hotspot suggestions or pattern recognition that highlights issues buried in raw numbers, but only if the experience stays grounded in human judgment. Councils need help, not automation that overrides local knowledge. So we’re taking deliberate steps to explore AI-assisted features once we’ve nailed coverage, accuracy, and transparency.
Filtering crash data surrouning Fitzroy North Primary School
Where We Are Now
CrashDash currently covers Victoria, New South Wales, Queensland, and Western Australia. South Australia is next in line this week, and we are moving steadily toward full national coverage before the PR launch. Each dataset we add forces us to refine the ingestion pipeline, tighten the mapping performance, and double-check the filters that make the interface feel trustworthy.
CrashDash sits within a wider CrowdLab roadmap. We are collaborating on new projects and experimenting with other community-focused tools that follow the same principles. The Pro tier for CrashDash is planned for early 2026. It will offer hotspot analytics, exportable reports, and account-level tools that larger councils and consultants need. The goal is to give power users the deeper analytics they have been asking for while keeping the core existing experience free and open to everyone. Think of it as the toolkit that lets a transport team walk into a budget meeting with the data-driven evidence they need already queued up.
Why It Matters
When local governments can see a clear, current picture of their crash landscape, they can deliver faster. Picture a transport engineer briefing councillors with a live map instead of a static PDF. Imagine a school principal pulling up the morning crash pattern before meeting with parents. Data-backed decisions become easier to justify. Road safety advocates can reference a shared source rather than relying on speculation. Communities gain a transparent view into what is happening near their schools, sports grounds, or bus corridors. Researchers and journalists can anchor their work in a national dataset rather than piecing together inconsistent fragments. This is about turning raw data into action. CrashDash opens the door for future CrowdLab products that follow the same philosophy.
Join In
If you work in transport, planning or advocacy, I built CrashDash so you don’t have to step through a state portal to get a clear answer.
Dive into the map and tell me what helps, what is missing, or where the next dataset should focus. Subscribing keeps you updated on the progress. A Pro tier waitlist will open soon, and I will share the link as soon as the landing experience is ready. In the meantime, reply if you want early access or have use cases I should prioritise.
— Anthony
Thanks for reading Anthony's Substack! Subscribe for free to receive new posts and support my work.
