Explore StreetLens: Simplified GIS Insights in 10 Seconds
For most of its life, StreetLens asked you to think like a GIS analyst before it would tell you anything. Fifteen layers, fourteen of them switched off, a legend showing raw quantile breaks, and the filters that actually matter — beds, baths, type, price — buried three clicks deep in a side panel.
The brief I wrote for this rebuild put it more bluntly than I would have in public:
"Getting value requires understanding GIS concepts. Hex layers occlude each other when multiple are enabled, the legend shows raw quantile breaks, and the useful filters are buried in a side panel. Mobile is effectively unusable."
The target was a single number: a subscriber with zero GIS knowledge gets a real insight within 10 seconds of the map loading — on desktop and on a phone — without power users losing anything they had.
That shipped. Here's the whole thing, new and old, so you know what you're holding.
The 10-second path
1. Pick a metric. One tap on the top bar — Price, Rent, Yield, Days on market, IRSAD — and that heatmap paints. Only one at a time, which fixes the single worst thing about the old map: stacked hexagons quietly hiding each other. Nine more sit under More, grouped Market / Area / Risk.
2. Read the legend, in English. It no longer just shows colours. It shows the distribution of the hexagons actually on your screen — a histogram, each colour band ending at its real breakpoint on a straight value axis, how many hexagons are mapped, how many have no data, the hexagon size in km², and how many properties the numbers are drawn from.
Kirwan, Townsville, houses, right now: modelled sale price $616k–$787k across 14 mapped hexagons, each 0.74 km², drawn from 8,076 properties. That's the whole basis of the picture, on screen, before you've clicked anything.
3. Click a hexagon. You get every metric HtAG holds for that cell in one card — Market, Area and Risk together — plus the listings inside it. Market hexagons and socio-economic hexagons are binned at different resolutions, so the card matches cells across resolutions and tells you the ground each block of numbers covers. Click a listing dot instead and the same card shows the property: price, beds/baths, type, agent, modelled value and rent, land size, floor area, last sale, days listed, link out to the listing.
4. Filter what's on the map. Type, Beds, Baths and Price are pills, always visible. Type is a single choice on purpose — the feed carries one row per (hexagon × property type), so "all types" would stack eight differently-valued hexagons on the same ground and colour whichever one drew last. Each type now shows how many hexagons it actually covers before you pick it, so landing on five hexagons of Farming isn't a surprise.
5. Work the list. The listings drawer holds the currently-filtered listings as rows — address, price (falling back to the modelled value, marked "est."), beds/baths, days listed — sortable by price, newest or days listed. Click a row and the map recentres on it and opens its card.
What's new beyond the basics
My Score. Pick up to three priorities, weight them, and the map repaints as a 0–100 match. Thirteen ingredients to choose from, each named for the direction that scores well — Low prices, High rents, Fast sales, Tightly held, Newer homes, Higher advantage, Less public housing, Big blocks, Fast leasing, Low risk. Inputs are min-max normalised across the mapped cells before weighting, so dollars, percentages and years can actually be blended. A cell is only scored where the inputs it has carry at least 60% of your weight — a hexagon with one of three ingredients shouldn't look as confident as a complete one.
Compare. Two metrics, one map, a swipe handle between them. Three presets — Price vs Yield, Price vs Flood risk, Rent vs Renters — or pick each side by hand.
Split view. Two panes for comparing two locations, and unlike stock kepler the second pane arrives pre-filled instead of blank. Keyboard: S.
Draw an area. Draw a polygon and you get a summary of what's inside it — median value for the active metric, hexagon count, listings inside with their median price, and the area in km². One button focuses the map on it, another gives you the whole suburb back.
Near here (Pro). From any hexagon or any listing: a 15-minute drive-time polygon, and what's nearby — schools, shops, transport, health, recreation, food — within 3 km. Then Only show listings inside, which clips the listings to the drive-time boundary and leaves the hexagons alone, so you keep the context you're reading the drive time against.
Copy link. Encodes your metric, your filters, your viewport, your compare sides and your score weights into the page URL. Send it and the recipient sees your map — not their own filters applied to your link. Come back later and your last view is restored, with a one-click reset.
Save image. Export the map with your legend in the picture, the drive-time and POI colour key alongside it, live-only controls stripped out and the HtAG mark top-right. It's built to go straight into a client report.
Advanced view. Nothing was taken away. One click bottom-left and the full kepler authoring UI is back — layer panel, colour configurators, the works, exactly as it was.
What sits underneath it
The interface is the new part. The data model it reads has been the quiet work:
Pre-aggregated H3 hexagons, computed server-side rather than binned in your browser. Market metrics at resolution 8 (~0.74 km² per hexagon), socio-economic and hazard at resolution 9 (~0.11 km²) — finer, because that data supports it.
Medians, not averages. One mansion or one granny flat can't move a cell. A hexagon needs at least two backing addresses to exist at all, and values beyond three standard deviations of their stratum are trimmed.
Effective resolution. Where a hexagon's local sample is too thin, it inherits its parent hexagon's median rather than showing you a hole. New estates get a sensible figure instead of blank ground.
Modelled vs actual, disclosed. Price and rent use a real recent transaction where one exists and an HtAG model estimate where it doesn't — and on a listing card it tells you which, with the as-at date.
Blank means unknown, never zero. Hexagons with no flood or bushfire exposure ship as blank so the risk layers paint only where risk exists, and the legend counts the no-data cells for you.
Listings refresh daily. Hex layers refresh monthly.
Suburb and LGA levels, across 7,000+ suburbs and 537 LGA dashboards.
And every metric now carries How this is modelled — what it shows, how it's built, how to read the map in front of you, and the caveats that matter. Written from the methodology, per metric, not one paragraph per group. Yield explains why a hexagon and a listing card disagree: the hexagon is a return on modelled value because an area has no single asking price, while a listing uses its own asking price.
What your plan opens
Public — Price.
Investor — Rent, Yield, Days on market, Affordability, Days on rental market, Hold period, Lot size, Property age, and the active listings feed.
Pro — IRSAD, Renters, Public housing, Flood risk, Bushfire risk, and Near here.
Locked metrics stay visible with the plan that unlocks them on the badge. Nothing is hidden from you — you can see exactly what you're not buying yet.
Open any suburb or LGA page and hit the StreetLens map → htag.com.au
Three things worth doing on your first pass:
Pick your own three priorities in My Score and see which pockets of your shortlist suburb actually match them.
Run Price vs Flood risk in Compare on a coastal or river suburb.
Click the hexagon your last purchase sits in and check the card against what you paid.