How to evaluate public art mapping software

Search for a public art request for proposals and you will find hundreds of calls for artists. Calls for software are rare. Almost no program runs a formal procurement to put its collection on a map. The decision gets made another way: someone asks a colleague what they use, a vendor gives a demo, a line item appears in next year's budget.

That is a reasonable way to buy. It just means the comparison happens in your head instead of on paper. This is the version on paper.

Start with the question that decides it

Not "what features does it have." Feature lists converge. Every option on this page will show you pins on a map.

Ask instead: who does this work in year three?

A collection is not static. Works get commissioned, repainted, relocated, and occasionally retired through deaccession. Artists change how they want to be credited. A mural on a private wall disappears when the building sells. Every option below is easy in month one. They diverge sharply around month thirty, when the person who set it up has moved on and nobody remembers the login.

Most stale public art maps were not badly built. They were built by someone who left.

What a real requirements list looks like

In 2024, Calgary Arts Development went to market for exactly this. Their RFP is public, and it is the most useful artifact in this whole category, because it is a real buyer writing down what they actually needed rather than a vendor describing what they happened to sell.

They asked for artwork information and images, multimedia content, search and filter, an interactive interface, location and navigation services, geofencing, accessibility options, self-guided tours including make-your-own, social sharing, offline access, and community-contributed content with moderation tools. On the staff side: a content management system for updates without a developer, WCAG 2.1 accessibility, documentation, staff training, and two years of ongoing support with an option to extend.

Full disclosure, since it matters for how you read that list: we answered that RFP, and the Calgary Public Art Guide is the result. Take the requirements as a useful template from a peer program. Take our name on it as a reason to check our claims against the others rather than take them on faith.

One thing that list quietly reveals. Calgary asked for software, training, and multi-year support in a single procurement. That is not a purchase. That is a service, bought through a form designed for purchases.

The six things you can actually buy

A national archive listing

The Public Art Archive, run by the nonprofit Creative West, is a free public database of roughly 30,000 artworks across many cities. Submitting is free. Paid tiers add an embeddable map and a collections management system built on CollectionSpace, museum registrar software.

As of July 2026 their published pricing is $295 a year for a standalone embeddable map, from $500 a year for their showcase and embed products, and from $1,500 for the CMS on collections under 150 works.

Choose it if your goal is presence and discovery rather than a designed experience, or your budget is genuinely zero. It is a real free tier, not a trial.

Know that your collection lives on their domain under their branding. A city's collection page carries the Public Art Archive logo and navigation, not the city's. There is no visitor tour or wayfinding feature.

Worth saying plainly: this is not an either-or. Submitting to a national archive costs nothing and adds a discovery surface you do not maintain. Several programs do that and run their own guide.

A tour app

STQRY (which now also owns OnCell, PocketSights, and My Tours), Otocast, izi.TRAVEL, and Bloomberg Connects build self-guided audio tours for museums, heritage sites, and cities. Several cities run their public art tours this way today. STQRY publishes pricing at $199 a month or $2,295 a year; Otocast quotes a setup fee plus a per-point-of-interest rate.

Choose it if the tour is the product. If what you want is twelve narrated stops downtown and you are not trying to steward a catalog, these are built for exactly that.

Know that they are audio-guide-first generalists. There is no artwork-specific record underneath, so attribution, materials, provenance, and condition have nowhere to live. If you already pay for one of these for a heritage tour, adding public art to it is genuinely cheaper than a second vendor.

A collections management system

PastPerfect, Gallery Systems TMS, Axiell, Argus, CatalogIt, and Artwork Archive are collections management tools for registrars. Accession numbers, loans, storage, conservation, provenance. TMS explicitly markets a public art module.

Choose it if you have a registrar, an indoor accessioned collection alongside the outdoor one, or documentation obligations that outrank public access.

Know that the public-facing side of most of these is a web portal, not a map-first experience. They are built for the person maintaining the record, not the person standing on the corner.

A GIS layer

Your city probably already owns ArcGIS. Saskatoon, Burnaby, Victoria, Tampa, and Tempe all publish public art this way, and the marginal cost is near zero.

Choose it if you need the data published and your GIS team is willing to own it.

Know that an open data layer is analyst-facing. It answers "where are the works" and not "what is this, who made it, and why should I walk four blocks to see it." It is a good input to a guide. It is rarely a good destination for a visitor. There is also a political dimension worth naming: if your GIS team built the current layer, replacing it is a conversation about their work, not just about software.

A map you build yourself

Google My Maps, a WordPress plugin, an Airtable base behind a page.

Choose it if you have fifteen works and no budget. Nothing on this page beats free at that scale, and a free map that exists is worth more than a paid one you are still scoping.

Know that it breaks somewhere around a hundred works, and it breaks quietly. Artist names get typed three ways. Two entries describe the same sculpture. Nobody notices until someone asks for a count.

A custom build

Hire a developer or an agency and own the result.

Choose it if you have a genuine reason to own the codebase, or capital funding that cannot be spent on a subscription.

Know that the build is the cheap part. STQRY's own writing cites around $100,000 for a private developer build plus thousands more for updates. Sydney Culture Walks launched in 2013 and shows its age. MIT had to rebuild its public art map outright when the original stopped working on phones. A one-off build starts aging on launch day, and the maintenance lands on you.

An operated guide

This is what we sell, so weigh it accordingly. Scoped onboarding to get the collection in, then a subscription under which we run it. We do not publish a price list, because the onboarding scope genuinely varies with the state of your data. If a vendor's pricing model is not obvious from their site, ask on the first call and expect a straight answer.

Choose it if the honest answer to "who does this in year three" is "nobody has time."

Know that you are buying a service, not a licence, and you should hold any vendor selling one to the data-ownership question below.

Questions worth asking any vendor

• Whose domain and whose brand does the public see? A subpath on a vendor's site is a different product from your own address.
• Can we export everything, today, without asking? Get the answer in writing. A catalog you cannot leave with is not yours.
• What happens when the person who set this up leaves? Ask who they train and what the handover looks like.
• Does it meet WCAG 2.1 AA, and where does it fall short? Every product has gaps. A vendor who names theirs is more trustworthy than one who claims none.
• How does pricing change as the collection grows? Per record, per tour, per point of interest, and flat annual are very different curves at 400 works.
• Can the public contribute, and who approves it? Community engagement without moderation is a reputational risk, not a feature.
• What did you ship in the last six months? A changelog is the cheapest honesty signal there is.

The year-three test

Picture the collection three years out. Forty new works. A different coordinator. A council asking what the program delivered.

Which of these options is still current in that picture, and who kept it that way? That answer will not come from a feature table. It comes from the pricing model, the support arrangement, and whether keeping it current is somebody's actual job.

Everything else is detail.

If the honest answer to the year-three question is that nobody has time, that is the specific problem we built Civic Collection to take off your desk. Book a demo and we will walk your collection through this list together.