Live axe-core inventory found 13 failing fg/bg pairs (3.12:1 coral text/
fills, 4.0-4.5 teal near-misses, 2.7 gold badge, 1.9 chart-colour text).
All replacement values validated ≥4.5:1 against every background they
sit on:
- --accent-coral-dark deepened to #b04a2e and used for coral text roles
(nav active, back/map links, kickers, chips) and coral fills under
white text (btn-primary, phase tabs, section-nav compare, chart chips);
new --accent-coral-darker #9c3f26 for their hovers. Decorative coral
(pins, bars, borders, focus ring) keeps the brand #e07256.
- --accent-teal darkened to #296f6f (rankings values, metric rows,
ofsted grades, eyebrows all pass).
- Ofsted green #3c8c3c→#2f7a2f; off-token gold #b8920e→--accent-gold-text.
- Gender-split pink text #b45778→#a04a68; admissions tip numerals to a
3.8:1 muted brown (28px display text).
- New CHART_TEXT_COLORS: AA-dark counterparts of the Chart.js series
palette for compare-card values (swatch dots keep true series colour).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
On ≤640px the floating Compare button becomes a 40px round glyph (+ / ✓,
aria-labelled) like the section-nav compare icon, instead of a full-width
text pill over the map. The OSM attribution moves to the map band's
top-left in preview so it no longer collides with the school name sliding
up under the fade; fullscreen keeps Leaflet's default bottom-right since
the zoom control occupies the top-left there.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Regression from d4d9ae5 (map-blended hero): .headerHasMap .actions was
absolutely positioned intending to float over the map, but its containing
block was .headerContent — made position:relative in the same commit — so
top:14px anchored it to the title block below the map. On mobile the H1
spans the full width and the legacy '.actions { width: 100% }' rule still
applied, stretching the glassy button across the school name and leaving
a ~6px legible sliver between it and the map fade.
Anchor .actions to .header by moving position:relative/z-index:3 from
.headerContent to .titleSection (the element that actually needs to sit
above the fade), and set width:auto on the floating variant so the mobile
full-width rule can't reach it. Applied to both detail views.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Recheck with 'B91 3DL' via the search box shows the geocoded radius
search works (distances, nearest-first, radius selector, map toggle).
Original P0.3 tested only a partial postcode, which silently falls back
to text search — refiled as P2.0 (silent mode switch on partial input).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Replace the faint middot separators on the characteristics line with
subtle borderless "quiet chips" so each attribute (type, age, faith,
gender) is its own scannable unit. Keep the coloured phase pill as the
one accent, and align every chip — including phase and provision tags —
to a single size/weight/radius (0.75rem / 600 / 4px).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The bare "3-11" range gave no unit. Add a display-only formatAgeRange
helper ("3-11" -> "Ages 3–11", en-dash) used in the primary/secondary
search rows and the secondary detail badge. The raw age_range field is
left untouched so sixth-form detection (.includes('18')) still works.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a per-step link on the primary and secondary admissions timelines
pointing to the relevant SchoolCompare tool (search, compare shortlist,
rankings) for that stage — from researching criteria through offer day
and appeals.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The report link hardcoded provider type 21, which 404s for every school
whose Ofsted provider type isn't 21 (nurseries -> 20, academies -> 23, etc.).
Use Ofsted's URN-based find-inspection-report URL, which redirects to the
correct provider page for any school, and rename the link to "Ofsted reports"
since it points to the school's Ofsted page (all reports), not one report.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The detail page reads overall_effectiveness from the API and showed
"Not rated" for ungraded-only schools, even though the list badge uses
the coalesced grade. Coalesce the ungraded fallback into the API's
overall_effectiveness so the detail page shows the same grade.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
int_ofsted_latest is only ref()'d inside a conditional block, so dbt
couldn't infer the edge and failed to compile dim_school. Add the
-- depends_on hint dbt recommends. No runtime behaviour change: the
adapter.get_relation guard still handles the pre-Ofsted-pipeline case.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Ungraded (Section 8) inspections don't assign a fresh grade — the export
only gives free text like "School remains Good". Parse that text into a
grade (remains Outstanding -> 1, remains Good -> 2, else null) and use it
as a last-resort fallback when no graded overall effectiveness exists.
Also retain schools that have only an ungraded inspection (no graded date)
by coalescing the inspection date, so ~8.5k previously-dropped schools now
carry a grade.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The blend gradient sat at z-index 2 — below Leaflet's tile pane (200) — so it
was painted behind the map and the band ended in a hard edge with no
diffusion. Raise the fade above the tile/overlay panes (450, still below the
marker pane so the pin stays crisp) and the whole-band open button above the
marker pane (800), and isolate the wrapper's stacking context so those raised
z-indexes don't leak out and outrank the header's Compare button.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace the header's at-a-glance stats row with a location map that sits
atop the hero and blends into the school title. The map is a static,
non-interactive preview (never traps page scroll) with a coral pin; the
whole band — or the inline "View on map ↗" link by the address — opens a
fullscreen, interactive map. Compare floats glassy over the band.
The separate "Location" section (and its nav item) is removed; the map now
lives only in the hero. Schools without lat/long render the header with no
map band, as before.
New: SchoolHeroMap (fullscreen wrapper, forwardRef open handle) +
LeafletHeroMapInner (minimal single-school map with interaction toggle).
Applied to both primary and secondary detail views; dead heroStats/tone/
mapContainer CSS removed (shared .heroStat* card classes kept).
Also drops the orphaned "Latest data" note and does not reintroduce an
Ofsted strip in the hero (it would duplicate the Ofsted section directly
below). Mockup kept at mockups/header-map-hero.html.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replicate the Autotrader pattern: pull the page-back control out of the
sticky section-nav bar and place a standalone "← Back" text link above the
header card, on the page background, where it scrolls away with the page.
It keeps the context-aware behaviour (router.back → /search fallback); the
secondary view gains the same fallback for parity (it previously used a bare
router.back() that dead-ends on deep links).
The pinned slot the Back button vacated now holds a "↑ Top" control that
smooth-scrolls to the top, so the sticky bar keeps a useful affordance.
Applies to both primary and secondary detail views.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A school inspected under the OEIF framework after September 2024 has an
inspection on record (ofsted_date set) but no overall_effectiveness grade,
since Ofsted no longer issues an overall judgement. buildOfstedListBadge
had no branch for this and fell through to "Not yet inspected", while the
detail page's hero chip correctly reported it as inspected — so the same
school read two contradictory ways.
Add an "Inspected · YYYY" branch that fires when an inspection is on record
(date or framework present) but no grade and not a Report Card, mirroring
the hero chip's fallback. Genuinely un-inspected schools (all Ofsted fields
null) still show "Not yet inspected". Add an .ofstedInspected badge style
(neutral slate) distinct from the grey pending state, and cover both cases
with tests.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The hero header rendered the Ofsted signal twice — once as a chip in the
strip and again as a tile in the at-a-glance scorecard — and an
Oversubscribed chip already covered by the First-choice tile's footnote.
Remove the chip strip on both primary and secondary detail views, leaving
the scorecard trio (Results · Ofsted · First-choice) as the single home
for the headline numbers. Widen the scorecard's render gate to hasHeroStats
so a school with an Ofsted rating but no results still shows its Ofsted
signal (previously carried by the chip strip). Drop the now-dead .heroChip
CSS, keeping the shared .tone-* tokens the scorecard's serif number uses.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
"How Hard to Get Into This School" presupposes difficulty, which misleads for the
many undersubscribed primaries (e.g. schools offering places to 100% of
first-choice applicants). Rename to the neutral "Admissions", matching the
secondary page and the section nav label; the demand figures and the
"Oversubscribed" chip convey the difficulty where it applies.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lh7Js5xSetKNzLVr9ArXLF
Reorders the detail-page sections so the most-sought content leads, based on
30-day section_nav_used analytics (nav clicks over-count buried sections, so a
low, high-demand section is a strong "surface me" signal):
Primary: Ofsted → SATs Results → Admissions → Pupils & Inclusion → History →
Phonics → What Parents Say → School Life → Location → Local Area → Finances
Secondary: Ofsted → GCSEs → Admissions → History → Parents → Wellbeing →
Location → Finances
Key moves: Admissions (23%, most-clicked) lifted from #6→#3; Pupils & Inclusion
(18%) from #7→#4; History (13%) off the bottom. Ofsted stays first (its low click
rate is positional — everyone reaches it first). Low-demand context sections
(Local Area, Finances) stay last. Brings the primary page in line with secondary.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lh7Js5xSetKNzLVr9ArXLF
The mobile breakpoint's display:none for .sectionNavAll / .sectionNavCompare was
declared before their base display rules, so equal-specificity cascade let the
base win and "All ▾" showed on mobile alongside the section menu. Move the
breakpoint swap block after the base declarations so the hide rules win.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lh7Js5xSetKNzLVr9ArXLF
Fixes the cramped mobile bar where Back, the carried Compare pill and All ▾ all
competed with the scrolling section links:
- On mobile the horizontal swipe strip collapses into a single "Section: <current> ▾"
button that opens the existing jump-to-section sheet — no fragile horizontal
swipe, and it doubles as a "you are here" indicator.
- The carried Compare CTA becomes a compact 38px icon on mobile (compare-arrows +
"add" badge; flips to a teal check when in the comparison), so it no longer
crowds the section control. Desktop keeps the labelled pill.
- Back is icon-only on mobile (label hidden), text on desktop.
- Desktop is unchanged: horizontal links + All ▾, full CTA in the hero.
Both the mobile section menu and the desktop All ▾ open the same sheet.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lh7Js5xSetKNzLVr9ArXLF
Reworks the sticky section bar on the school detail page so mobile users can
perceive the page's breadth and get back to their results without hunting:
- Dock the bar directly under the global header (top: 56px mobile / 64px desktop)
so it's reachable immediately instead of buried below the tall hero.
- Pin "← Back" so it never scrolls off; only the section links scroll, keeping
the existing right-edge fade. Back uses router.back() with a /search fallback
so deep-links never dead-end.
- Carry the hero's "Add to Compare" CTA into the bar as a compact pill once the
hero button scrolls out of view, so the conversion action survives.
- Add an "All ▾" menu (dropdown on desktop, bottom sheet on mobile) listing every
section with the active one ticked — full breadth without horizontal scrolling.
Reuses the site's existing pill / coral / edge-fade / sheet vocabulary rather
than introducing a new navigation paradigm.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lh7Js5xSetKNzLVr9ArXLF
The "How Hard to Get Into This School" this-year view rendered four
question/answer rows stretched to fill the trend chart's height, leaving
tall gaps and a long horizontal jump from a small muted question to its
number. Replace with a 2×2 grid of stat tiles — each number paired with
its label directly beneath — so the panel reads quickly and fills the
reserved height without artificial spacing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A first-choice offer rate sitting on the y-max ceiling (100%) was drawn
flush against the top of the plot area, clipping the point marker. Add 8px
top layout padding and clip:false so ceiling values render in full, without
introducing a misleading >100% tick.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Brings the AdmissionsTrendChart (Chart.js) fix into main; the earlier PR
merged only the initial SVG version, which rendered with oversized,
overlapping labels.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The hand-rolled SVG sparkline used px font sizes inside a 520-wide viewBox
that stretched to the full card width, so labels ballooned ~4x and collided —
and with 10+ years of real data the per-point labels and year ticks
overlapped badly, while the oversized chart stretched the "this year" view.
Replace it with a Chart.js line chart (AdmissionsTrendChart) in a fixed
200px wrapper, matching PerformanceChart: responsive px fonts, auto-skipping
x ticks, auto-scaled y-axis clamped to 0-100, emphasised latest point.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The school detail page only showed the latest admissions year. We store
every year, which is more decision-relevant for parents (the trend and its
consistency matter more than a single noisy year).
Backend now returns the full admissions_history (oldest first) alongside the
existing latest-year object. The primary SchoolDetailView gains a header
toggle ("This year | N-year trend") that swaps the Q&A for an SVG sparkline
of the first-choice offer rate. The toggle only appears when >=2 years carry
an offer rate; otherwise it falls back to the single-year card. Both views
share one CSS-grid cell so switching causes no layout shift.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The standalone dbt Fusion binary (dbt-core 2.x) on PATH shadows the
pip-installed classic dbt-postgres ~=1.10 and rejects the Postgres
adapter (dbt1005), breaking every DAG's dbt_build task. Invoke dbt via
`python -m dbt.cli.main` in the DAGs and the Dockerfile dbt deps step so
the classic Postgres-capable engine is always used regardless of PATH.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The "Peaked at X%" hint didn't say which metric it referenced, leaving
parents to guess. Both branches now lead with "Reading, Writing & Maths".
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Frontend
- Dynamic-import Chart.js components on detail/compare views so Chart.js
no longer ships in initial JS.
- Drop force-dynamic on home, compare, rankings so internal data fetches
reuse Next.js's per-call revalidate cache.
- Switch /school/[slug] to ISR with a 7-day revalidate window (school
data updates annually).
- Preconnect to analytics + postcodes.io; remove redundant defer on the
Umami Script tag (afterInteractive already covers it).
- Bump images.minimumCacheTTL to 1 year.
- Extract HowItWorks and Editorial sections as server components passed
to HomeView via slot props so their JSX stays out of the client bundle.
Backend
- Add GZipMiddleware (min 512 bytes).
- Add CacheAndETagMiddleware: per-path Cache-Control with long s-maxage
+ stale-while-revalidate, ETag generation, and 304 on If-None-Match.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds data-performance="true" to the Umami script tag. From v3.1.0+,
this hooks PerformanceObserver and posts LCP, INP, CLS, FCP and TTFB
to the collect endpoint with Google's rating buckets pre-applied.
The Umami dashboard then surfaces them in its built-in Performance
tab with p50/p75/p95 percentiles and per-page / per-device breakdowns
— no custom event taxonomy needed.
Requires the analytics instance to be on Umami ≥ v3.1.0. The
attribute is a no-op on older versions, so no risk to ship before
upgrading the dashboard.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add lib/analytics.ts with a single typed track() wrapper. SSR-safe,
never throws, no-ops when Umami isn't loaded. Event names form a
fixed union so refactors stay safe.
14 events wired:
Discovery (3)
search_submitted FilterBar submit + near_me path
near_me_used all geolocation outcomes
empty_results search returns 0 schools
Engagement (5)
school_viewed SchoolDetail + Secondary on mount, with
urn / phase / local_authority / from
section_nav_used section-nav links on both detail views
chart_metric_changed mobile chart chip switch
metric_compared_in_rankings rankings metric dropdown
external_link_clicked Ofsted / school website / DfE (declarative
data-umami-event attributes)
Conversion (5)
compare_school_added search/rankings/detail/compare sources
compare_school_removed detail toggle and compare page
compare_viewed once per session when there's a selection
(school_count, phase_mix)
compare_metric_changed compare page metric dropdown
compare_shared native sheet vs clipboard distinguished
Operational (1)
api_error caught in handleResponse, includes
endpoint / status / route
Suggested Goals to configure in the Umami dashboard for the funnel
report: search_submitted → school_viewed → compare_school_added →
compare_viewed → compare_shared.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
.chartWrapper had height:100% but its parent .chartOuter (a flex
column) had no explicit height — so the canvas couldn't resolve a
real height and fell back to Chart.js's small default, leaving a
big empty band between the plot and the "View raw year-by-year data"
disclosure on desktop.
Give .chartOuter height:100% so it fills the 280px .chartContainer,
and switch .chartWrapper to flex:1 1 auto / min-height:0 so the
canvas fills whatever space remains after the (primary-only) trend
banner. Mobile's explicit .chartWrapper height:220px still wins
inside the responsive override.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The expected-standard ranking groups too many schools at or near
100% — the leaderboard isn't useful when the top 30+ schools all
share the same value. Switch the default primary metric to the
higher (above-expected) standard, which discriminates between
schools much more clearly at the top end. Secondary still defaults
to Attainment 8. Users can still pick either via the dropdown.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The previous commit added the chip strip and subtitle inside the
.chartOuter, but .chartContainer on the parent SchoolDetailView still
had a fixed height: 220px. With the new content stacking above the
canvas the chart + COVID footnote overflowed past 220px, and the
<details> "View raw year-by-year data" disclosure (rendered just
after the container) landed on top of the plot. It also pushed the
COVID footnote out of the card onto the page background.
- .chartContainer at ≤640px now flows naturally (height: auto)
- .chartWrapper at ≤640px gets an explicit 220px height so the canvas
itself still has a known size for Chart.js to render into.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The Results Over Time chart on the school detail page was overcrowded
on phones — 5-item legend wrapping over the plot area, a hidden right
y-axis still rendered for the (collapsed) progress series, fixed
0–100 percentage scale that flattened all variation, angled x-axis
labels eating vertical space.
At ≤640px the chart now renders one metric at a time, selected via a
chip strip above the plot. No more legend. No more dual y-axis. The
y-axis auto-tightens around the actual data range so variation is
visible (a school sitting in the 70–90% band now uses a 65–95 axis
instead of squashing onto a 0–100 line). A small subtitle above the
chips sets the subject context ("KS2 SATs · Reading, Writing & Maths"
or "GCSE results · Year 11") so chip labels can describe the *view*
rather than re-stating the subject.
Chip labels are spelled out in parent-facing language — no internal
shorthand like "RWM":
Primary (KS2): At expected level (default)
Above expected level
Pupil progress (shows 3 series + mini-legend)
Secondary (KS4): Attainment 8 (default)
English & Maths grade 4+
Progress 8
Chips disable themselves (greyed, with a "No data for this school"
title) when the underlying series has no data points. Desktop
behaviour is unchanged — the full multi-series chart with dual y-axis
still renders >640px.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
MOB-08: Search list pagination is already implemented (page_size 50,
"Load more schools" button + count) — no change needed; ticket closed.
MOB-10: Rather than build a full bottom-sheet filter modal (large
change, modal/focus-trap/scroll-lock infra), promote the existing
"Advanced" toggle to a coral pill labeled "Filters (n)" whenever
dropdown filters are applied. Users now see at a glance that the list
is being narrowed; the inline accordion remains the disclosure
mechanism. Adds aria-expanded for screen readers.
MOB-23: Add MOBILE.md at the repo root with the 360 px design baseline,
acceptance checks for any UI PR (no horizontal overflow, ≥44px tap
targets, no <11px visible text, iOS Chrome parity, safe-area-inset,
dvh), and the established component patterns. Playwright regression
test deferred — adding the dep for one test is heavier than the
current value warrants; documented as a future option.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
tudor