fix: make globe country selection reliable

Author: Oussama

data/build_aggregates.py

@@ -340,6 +340,7 @@ def artist_story(
     last_decade: int | None,
     work_title: str | None,
     work_year: int | None,
+    score: float,
 ) -> str:
     """Build a factual collection note without inventing biography."""
     country_phrase = (
@@ -360,12 +361,51 @@ def artist_story(
             work += f" ({work_year})"
         work += "."
     return (
-        f"In this dataset, {artist_name} is the most represented artist credited to {country_phrase}, "
-        f"with {n_works:,} credited works. MoMA's records place this artist mostly in {top_medium}, "
+        f"In this dataset, {artist_name} has the highest collection prominence score for {country_phrase} "
+        f"({score:.1f}/100), with {n_works:,} credited works. MoMA's records place this artist mostly in {top_medium}, "
         f"with works appearing {span}.{work}"
     )
 
 
+def prominence_table(group: pd.DataFrame) -> pd.DataFrame:
+    """Rank artists within one country using a transparent collection-based proxy."""
+    stats = (
+        group.groupby("artist_id")
+        .agg(
+            artist_name=("artist_name", "first"),
+            n_works=("artwork_id", "count"),
+            n_departments=("department", "nunique"),
+            n_mediums=("medium_group", "nunique"),
+            first_decade=("decade", "min"),
+            last_decade=("decade", "max"),
+            cataloged_share=("cataloged", lambda values: float((values == "Y").mean())),
+            n_on_view=("on_view", lambda values: int(values.notna().sum())),
+            has_wiki=("has_wiki", "max"),
+            has_ulan=("has_ulan", "max"),
+        )
+        .reset_index()
+    )
+    stats["decade_span"] = (stats["last_decade"] - stats["first_decade"]).fillna(0).clip(lower=0)
+    max_works = max(1, int(stats["n_works"].max()))
+    max_departments = max(1, int(stats["n_departments"].max()))
+    max_mediums = max(1, int(stats["n_mediums"].max()))
+    max_on_view = max(1, int(stats["n_on_view"].max()))
+    stats["prominence_score"] = (
+        55 * stats["n_works"].map(lambda value: math.log1p(value) / math.log1p(max_works))
+        + 12 * (stats["n_departments"] / max_departments)
+        + 8 * (stats["n_mediums"] / max_mediums)
+        + 8 * (stats["decade_span"].clip(upper=160) / 160)
+        + 7 * stats["cataloged_share"]
+        + 5 * stats["has_wiki"].astype(int)
+        + 3 * stats["has_ulan"].astype(int)
+        + 2 * (stats["n_on_view"] / max_on_view)
+    ).round(2)
+    return stats.sort_values(
+        ["prominence_score", "n_works", "artist_name"],
+        ascending=[False, False, True],
+    )
+
+
 def load_lookup_files() -> tuple[dict[str, str], dict[str, str]]:
     with (DATA_DIR / "nationality_to_iso3.json").open("r", encoding="utf-8") as handle:
         raw_nationality_to_iso = json.load(handle)
@@ -410,6 +450,8 @@ def build_credit_rows(artworks: pd.DataFrame, nationality_to_iso: dict[str, str]
                     "classification": clean_token(getattr(row, "Classification")) or "Unknown",
                     "medium": clean_token(getattr(row, "Medium")) or "Unknown",
                     "medium_group": getattr(row, "medium_group"),
+                    "cataloged": clean_token(getattr(row, "Cataloged")),
+                    "on_view": clean_token(getattr(row, "OnView")),
                     "nationality": nationality,
                     "iso3": iso3,
                     "country_name": ISO3_TO_COUNTRY.get(iso3, iso3) if iso3 else None,
@@ -456,6 +498,10 @@ def main() -> None:
 
     credit_rows = build_credit_rows(artworks, nationality_to_iso, regions)
     credits = pd.DataFrame(credit_rows)
+    wiki_lookup = artists_raw.set_index("ConstituentID")["Wiki QID"].notna().to_dict()
+    ulan_lookup = artists_raw.set_index("ConstituentID")["ULAN"].notna().to_dict()
+    credits["has_wiki"] = credits["artist_id"].map(wiki_lookup).fillna(False)
+    credits["has_ulan"] = credits["artist_id"].map(ulan_lookup).fillna(False)
     n_total = len(artworks)
     n_credits = len(credits)
 
@@ -514,7 +560,9 @@ def main() -> None:
         medium_counter = Counter(group["medium_group"])
         sorted_group = group.sort_values(["artist_name", "title", "year"], na_position="last")
         sample = sorted_group.iloc[seeded_index(str(iso3), len(sorted_group))]
-        featured_artist_id = int(group["artist_id"].value_counts().sort_values(ascending=False).index[0])
+        ranked_artists = prominence_table(group)
+        featured_rank = ranked_artists.iloc[0]
+        featured_artist_id = int(featured_rank["artist_id"])
         featured_group = group[group["artist_id"] == featured_artist_id].sort_values(["year", "title"], na_position="last")
         featured_sample = featured_group.iloc[0]
         featured_medium = Counter(featured_group["medium_group"]).most_common(1)[0][0]
@@ -539,6 +587,17 @@ def main() -> None:
                 "featured_artist_id": featured_artist_id,
                 "featured_artist": featured_sample["artist_name"],
                 "featured_artist_n_works": int(len(featured_group)),
+                "featured_artist_score": float(featured_rank["prominence_score"]),
+                "featured_artist_score_method": (
+                    "Collection prominence score: 55% log work count, 12% department breadth, "
+                    "8% medium breadth, 8% dated decade span, 7% cataloged share, "
+                    "5% Wiki QID, 3% ULAN, 2% currently-on-view records."
+                ),
+                "featured_artist_n_departments": int(featured_rank["n_departments"]),
+                "featured_artist_n_mediums": int(featured_rank["n_mediums"]),
+                "featured_artist_decade_span": int(json_ready(featured_rank["decade_span"]) or 0),
+                "featured_artist_has_wiki": bool(featured_rank["has_wiki"]),
+                "featured_artist_has_ulan": bool(featured_rank["has_ulan"]),
                 "featured_artist_lifespan": lifespan(
                     featured_sample["year_birth"],
                     featured_sample["year_death"],
@@ -558,6 +617,7 @@ def main() -> None:
                     featured_last,
                     featured_sample["title"],
                     featured_work_year,
+                    float(featured_rank["prominence_score"]),
                 ),
             }
         )

data/build_report.txt

@@ -1,5 +1,5 @@
 DArts aggregate build report
-generated_at: 2026-05-08T21:20:30+00:00
+generated_at: 2026-05-08T22:24:24+00:00
 
 raw_artworks: 160,248
 cleaned_artworks: 144,149
@@ -42,7 +42,7 @@ top_unmapped_nationality_labels:
 outputs:
   artist_index.json: 3,548,506 bytes
   country_by_decade.json: 50,233 bytes
-  country_summary.json: 113,489 bytes
+  country_summary.json: 162,166 bytes
   gender_by_decade.json: 3,753 bytes
   gender_by_decade_country.json: 79,461 bytes
   gender_by_decade_department.json: 21,341 bytes

website/public/data/country_summary.json

@@ -1 +1 @@
-{"n_artworks_total":144149,"n_artworks_dated":141884,"n_artist_credits":160035,"n_artists_total":11879,"year_min":1768,"year_max":2026,"generated_at":"2026-05-08T21:20:30+00:00","total_artworks":144149,"total_artists":11879,"unique_nationalities":129,"top_department":"Drawings & Prints"}
+{"n_artworks_total":144149,"n_artworks_dated":141884,"n_artist_credits":160035,"n_artists_total":11879,"year_min":1768,"year_max":2026,"generated_at":"2026-05-08T22:24:24+00:00","total_artworks":144149,"total_artists":11879,"unique_nationalities":129,"top_department":"Drawings & Prints"}

website/public/data/summary.json

@@ -17,6 +17,8 @@
   let dragging = $state(false);
   let dragStart = null;
   let rotateStart = null;
+  let pointerDownCountryIndex = null;
+  let pointerMoved = false;
   let frame = 0;
   let reduceMotion = false;
 
@@ -101,18 +103,27 @@
     dragging = true;
     dragStart = [event.clientX, event.clientY];
     rotateStart = rotation;
+    pointerMoved = false;
+    pointerDownCountryIndex = event.target?.dataset?.countryIndex ?? null;
     event.currentTarget.setPointerCapture(event.pointerId);
   }
 
   function onPointerMove(event) {
     if (!dragging || !dragStart || !rotateStart) return;
     const dx = event.clientX - dragStart[0];
     const dy = event.clientY - dragStart[1];
+    if (Math.hypot(dx, dy) > 4) pointerMoved = true;
     rotation = [rotateStart[0] + dx * 0.35, Math.max(-65, Math.min(65, rotateStart[1] - dy * 0.35))];
   }
 
-  function onPointerUp() {
+  function onPointerUp(event) {
+    const countryIndex = pointerDownCountryIndex;
+    const shouldSelect = dragging && !pointerMoved && countryIndex !== null;
     dragging = false;
+    pointerDownCountryIndex = null;
+    if (event.currentTarget.hasPointerCapture?.(event.pointerId))
+      event.currentTarget.releasePointerCapture(event.pointerId);
+    if (shouldSelect) centerCountry(countries[Number(countryIndex)]);
   }
 
   onMount(async () => {
@@ -150,16 +161,17 @@
       onpointerup={onPointerUp}
       onpointercancel={onPointerUp}
     >
-      {#each countries as country}
+      {#each countries as country, index}
         {@const iso3 = worldAtlasIdToIso3[String(country.id).padStart(3, '0')]}
         <path
           d={path(country)}
+          data-country-index={index}
+          data-iso3={iso3}
           fill={fillFor(iso3)}
           class:selected={selectedCountry === iso3}
           class:clickable={Boolean(iso3 && countByIso[iso3])}
           onpointermove={(event) => showTooltip(event, country)}
           onpointerleave={hideTooltip}
-          onclick={() => centerCountry(country)}
           onkeydown={(event) => {
             if (event.key === 'Enter' || event.key === ' ') {
               event.preventDefault();

website/src/lib/charts/Globe.svelte

@@ -28,6 +28,7 @@
     selectedSummary?.featured_story ??
       `${selectedSummary?.sample_artist} appears in the selected country's records with the representative work ${selectedSummary?.sample_work_title}.`,
   );
+  let selectedScore = $derived(selectedSummary?.featured_artist_score ?? null);
 
   function computeTopThreeShare(rows, decadeRange) {
     const counts = {};
@@ -116,7 +117,7 @@
           </div>
         </dl>
         <article class="artist-label">
-          <p class="eyebrow">Most represented artist in this slice</p>
+          <p class="eyebrow">Collection prominence pick</p>
           <div class="artist-heading">
             <span class="monogram" aria-hidden="true">{artistInitials(selectedArtistName)}</span>
             <div>
@@ -133,14 +134,38 @@
             </div>
           </div>
           <p class="story">{selectedStory}</p>
+          {#if selectedScore !== null}
+            <dl class="score-card" aria-label="Collection prominence score components">
+              <div>
+                <dt>Score</dt>
+                <dd>{selectedScore.toFixed(1)}/100</dd>
+              </div>
+              <div>
+                <dt>Works</dt>
+                <dd>{selectedSummary.featured_artist_n_works.toLocaleString()}</dd>
+              </div>
+              <div>
+                <dt>Breadth</dt>
+                <dd>
+                  {selectedSummary.featured_artist_n_departments} departments · {selectedSummary.featured_artist_n_mediums}
+                  media
+                </dd>
+              </div>
+            </dl>
+          {/if}
           <p class="representative">
             <span>Representative record</span>
             <em>{selectedWorkTitle}</em>
             {#if selectedWorkYear}
               <span>{selectedWorkYear}</span>
             {/if}
           </p>
-          <p class="caveat">This is a collection-count proxy, not a fame ranking.</p>
+          <p class="caveat">
+            This is a transparent collection-based proxy, not a true fame ranking.
+            {#if selectedSummary.featured_artist_score_method}
+              {selectedSummary.featured_artist_score_method}
+            {/if}
+          </p>
           {#if selectedArtistUrl}
             <a class="moma-link" href={selectedArtistUrl} target="_blank" rel="noopener">Open artist on MoMA</a>
           {/if}

website/src/lib/scenes/Scene2Globe.svelte

@@ -207,6 +207,34 @@
   line-height: var(--type-body-line);
 }
 
+.artist-label .score-card {
+  display: grid;
+  grid-template-columns: repeat(3, minmax(0, 1fr));
+  gap: var(--space-1);
+  margin: var(--space-2) 0 0;
+  padding: var(--space-1) 0 0;
+  border-top: 1px solid color-mix(in srgb, var(--fg-on-dark-mute) 28%, transparent);
+}
+
+.artist-label .score-card div {
+  display: block;
+}
+
+.artist-label .score-card dt {
+  color: var(--fg-on-dark-mute);
+  font-family: var(--font-ui);
+  font-size: var(--type-small-size);
+  text-transform: uppercase;
+}
+
+.artist-label .score-card dd {
+  margin: 0.125rem 0 0;
+  color: var(--fg-on-dark-strong);
+  font-family: var(--font-mono);
+  font-size: var(--type-small-size);
+  font-variant-numeric: tabular-nums;
+}
+
 .artist-label .representative {
   display: grid;
   gap: 0.125rem;
@@ -289,6 +317,12 @@
     top: auto;
     bottom: var(--space-2);
     width: auto;
+    max-height: calc(100svh - var(--space-4));
+    overflow: auto;
+  }
+
+  .artist-label .score-card {
+    grid-template-columns: 1fr;
   }
 }