HTTPArchive · max-ostapenko · Jan 15, 2026 · Jan 12, 2026 · Jan 13, 2026 · Jan 14, 2026
diff --git a/sql/2025/privacy/bounce_domains_top.sql b/sql/2025/privacy/bounce_domains_top.sql
@@ -73,6 +73,6 @@ FROM bounce_sequences
   ANY_VALUE(websites_pct) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop, cnt_mobile AS mobile_count, cnt_desktop AS desktop_count
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop, cnt_mobile AS mobile_count, cnt_desktop AS desktop_count
 |> ORDER BY COALESCE(mobile_count, 0) + COALESCE(desktop_count, 0) DESC
 |> LIMIT 100
diff --git a/sql/2025/privacy/client_hints_top.sql b/sql/2025/privacy/client_hints_top.sql
@@ -40,5 +40,5 @@ GROUP BY client, value
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/client_hints_usage.sql b/sql/2025/privacy/client_hints_usage.sql
@@ -51,4 +51,4 @@ GROUP BY all_accept_ch.client
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
diff --git a/sql/2025/privacy/cookie_domains_third_party_top.sql b/sql/2025/privacy/cookie_domains_third_party_top.sql
@@ -15,8 +15,6 @@ GROUP BY client, cookie_domain
   ANY_VALUE(pct_domains) AS pct_domains
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME
-pct_domains_mobile AS mobile,
-pct_domains_desktop AS desktop
+|> RENAME pct_domains_mobile AS Mobile, pct_domains_desktop AS Desktop
 |> ORDER BY COALESCE(domain_count_mobile, 0) + COALESCE(domain_count_desktop, 0) DESC
 |> LIMIT 1000
diff --git a/sql/2025/privacy/cookies_first_party_top.sql b/sql/2025/privacy/cookies_first_party_top.sql
@@ -20,8 +20,6 @@ GROUP BY client, cookie_name
   ANY_VALUE(pct_domains) AS pct_domains
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME
-pct_domains_mobile AS mobile,
-pct_domains_desktop AS desktop
+|> RENAME pct_domains_mobile AS Mobile, pct_domains_desktop AS Desktop
 |> ORDER BY COALESCE(domain_count_mobile, 0) + COALESCE(domain_count_desktop, 0) DESC
 |> LIMIT 1000
diff --git a/sql/2025/privacy/cookies_third_party_top.sql b/sql/2025/privacy/cookies_third_party_top.sql
@@ -16,8 +16,6 @@ GROUP BY client, cookie_details
   ANY_VALUE(pct_domains) AS pct_domains
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME
-pct_domains_mobile AS mobile,
-pct_domains_desktop AS desktop
+|> RENAME pct_domains_mobile AS Mobile, pct_domains_desktop AS Desktop
 |> ORDER BY COALESCE(domain_count_mobile, 0) + COALESCE(domain_count_desktop, 0) DESC
 |> LIMIT 1000
diff --git a/sql/2025/privacy/dnt_usage.sql b/sql/2025/privacy/dnt_usage.sql
@@ -15,5 +15,5 @@ FROM `httparchive.blink_features.usage`
   ANY_VALUE(pct_urls) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY rank ASC
diff --git a/sql/2025/privacy/fingerprinting_top.sql b/sql/2025/privacy/fingerprinting_top.sql
@@ -27,5 +27,5 @@ GROUP BY client, technology.technology
   ANY_VALUE(websites_pct) AS websites_pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME websites_pct_mobile AS mobile, websites_pct_desktop AS desktop
+|> RENAME websites_pct_mobile AS Mobile, websites_pct_desktop AS Desktop
 |> ORDER BY websites_count_mobile + websites_count_desktop DESC
diff --git a/sql/2025/privacy/iab_tcf_v2_cmps_top.sql b/sql/2025/privacy/iab_tcf_v2_cmps_top.sql
@@ -17,5 +17,5 @@ GROUP BY client, cmpId
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/iab_tcf_v2_countries_top.sql b/sql/2025/privacy/iab_tcf_v2_countries_top.sql
@@ -35,5 +35,5 @@ FROM base_data
   ANY_VALUE(pct_of_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/iab_usage.sql b/sql/2025/privacy/iab_usage.sql
@@ -53,5 +53,5 @@ FROM aggregated,
   ANY_VALUE(number_of_websites) AS websites_count
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY websites_count_desktop + websites_count_mobile DESC
diff --git a/sql/2025/privacy/iab_usp_strings_top.sql b/sql/2025/privacy/iab_usp_strings_top.sql
@@ -24,5 +24,5 @@ FROM `httparchive.crawl.pages`
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/referrer_policy_top.sql b/sql/2025/privacy/referrer_policy_top.sql
@@ -56,5 +56,5 @@ FROM referrer_policy_custom_metrics
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/referrer_policy_usage.sql b/sql/2025/privacy/referrer_policy_usage.sql
@@ -54,5 +54,5 @@ FROM aggregated,
   ANY_VALUE(pct) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY mobile + desktop DESC
diff --git a/sql/2025/privacy/related_origin_trials_top.sql b/sql/2025/privacy/related_origin_trials_top.sql
@@ -110,5 +110,5 @@ FROM aggregated
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/tracker_categories_top.sql b/sql/2025/privacy/tracker_categories_top.sql
@@ -29,5 +29,5 @@ FROM `httparchive.crawl.pages`,
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/tracker_distribution.sql b/sql/2025/privacy/tracker_distribution.sql
@@ -50,5 +50,5 @@ FROM tracker_counts
   ANY_VALUE(ccdf) AS ccdf
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME ccdf_mobile AS mobile, ccdf_desktop AS desktop
+|> RENAME ccdf_mobile AS Mobile, ccdf_desktop AS Desktop
 |> ORDER BY number_of_trackers
diff --git a/sql/2025/privacy/tracker_technologies_top.sql b/sql/2025/privacy/tracker_technologies_top.sql
@@ -29,5 +29,5 @@ FROM `httparchive.crawl.pages`,
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/whotracksme_categories_top.sql b/sql/2025/privacy/whotracksme_categories_top.sql
@@ -62,5 +62,5 @@ FROM aggregated
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/sql/2025/privacy/whotracksme_trackers_top.sql b/sql/2025/privacy/whotracksme_trackers_top.sql
@@ -33,5 +33,5 @@ FROM `httparchive.crawl.requests`
   ANY_VALUE(pct_websites) AS pct
   FOR client IN ('desktop', 'mobile')
 )
-|> RENAME pct_mobile AS mobile, pct_desktop AS desktop
+|> RENAME pct_mobile AS Mobile, pct_desktop AS Desktop
 |> ORDER BY COALESCE(websites_count_desktop, 0) + COALESCE(websites_count_mobile, 0) DESC
diff --git a/src/tools/scripts/chart-adjustments.ipynb → sql/util/chart-adjustments.ipynb b/src/tools/scripts/chart-adjustments.ipynb → sql/util/chart-adjustments.ipynb
@@ -12,15 +12,39 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Authenticate the user\n",
+    "import sys\n",
+    "\n",
+    "if 'google.colab' in sys.modules:\n",
+    "    from google.colab import auth\n",
+    "\n",
+    "    auth.authenticate_user()\n",
+    "    credentials = auth.get_user_credentials()\n",
+    "else:\n",
+    "    import google.auth\n",
+    "\n",
+    "    SCOPES = [\n",
+    "        'https://www.googleapis.com/auth/spreadsheets'\n",
+    "    ]\n",
+    "    credentials, project = google.auth.default(scopes=SCOPES)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
    "metadata": {
     "id": "cOcbpC6qRou_"
    },
    "outputs": [],
    "source": [
-    "from google.colab import auth\n",
     "from googleapiclient.discovery import build\n",
     "\n",
+    "sheets_service = build('sheets', 'v4', cache_discovery=False, credentials=credentials)\n",
+    "\n",
     "def update_chart_size(spreadsheet_id, is_dry_run=False, target_width=600, target_height=371):\n",
     "    response = sheets_service.spreadsheets().get(spreadsheetId=spreadsheet_id, includeGridData=False).execute()\n",
     "    sheets = response.get('sheets', [])\n",
@@ -54,26 +78,22 @@
     "                print(f\"\"\"sheet: {sheet['properties']['title']},\n",
     "chart: {chart['spec']['title']},\n",
     "dimensions: {chart['position']['overlayPosition']['widthPixels']} x {chart['position']['overlayPosition'].get('heightPixels', 'N/A')}\n",
-    "                \"\"\")\n",
-    "\n",
-    "# Authenticate the user\n",
-    "auth.authenticate_user()\n",
-    "sheets_service = build('sheets', 'v4', cache_discovery=False)"
+    "                \"\"\")"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "metadata": {
     "id": "vp1izUBSLxp9"
    },
    "outputs": [],
    "source": [
     "# Replace this with the ID of your Google Sheets file\n",
-    "SPREADSHEET_ID = '18r8cT6x9lPdM-rXvXjsqx84W7ZDdTDYGD59xr0UGOwg'\n",
+    "SPREADSHEET_ID = '1Svyw40Th7VbigX6lpR1lb1WXwTUVKZWrK7O2YELrml4'\n",
     "\n",
     "# Call the function to update the chart width\n",
-    "update_chart_size(SPREADSHEET_ID, target_height=None, is_dry_run=True)"
+    "update_chart_size(SPREADSHEET_ID, target_height=None, is_dry_run=False)"
    ]
   }
  ],
@@ -85,11 +105,21 @@
    "provenance": []
   },
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": ".venv",
+   "language": "python",
    "name": "python3"
   },
   "language_info": {
-   "name": "python"
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.14.2"
   }
  },
  "nbformat": 4,

diff --git a/sql/util/generate_figure_markup.py b/sql/util/generate_figure_markup.py
@@ -0,0 +1,80 @@
+import re
+import os
+from googleapiclient.discovery import build  # pylint: disable=import-error
+import google.auth  # pylint: disable=import-error
+
+# Configuration
+SPREADSHEET_ID = '1Svyw40Th7VbigX6lpR1lb1WXwTUVKZWrK7O2YELrml4'
+PUBCHART_ID = '2PACX-1vRC5wrzy5NEsWNHn9w38RLsMURRScnP4jgjO1mDiVhsfFCY55tujlTUZhUaEWzmPtJza0QA7w8S4uK5'
+SQL_DIR = '../2025/privacy'  # Relative to this script's location
+
+SCOPES = ['https://www.googleapis.com/auth/spreadsheets']
+
+
+def get_sql_to_sheet_map(sql_dir):
+    mapping = {}
+    if not os.path.exists(sql_dir):
+        print(f"Directory not found: {sql_dir}")
+        return mapping
+    for filename in os.listdir(sql_dir):
+        if filename.endswith(".sql"):
+            # Generate sheet name from filename using the regex:
+            # re.sub(r'(\.sql|[^a-zA-Z0-9]+)', ' ', filename).strip().title()
+            sheet_name = re.sub(r'(\.sql|[^a-zA-Z0-9]+)', ' ', filename).strip().title()
+            mapping[sheet_name] = filename
+    return mapping
+
+
+def generate_figure_markup(spreadsheet_id, sql_dir):
+    try:
+        credentials, project = google.auth.default(scopes=SCOPES)
+        sheets_service = build('sheets', 'v4', cache_discovery=False, credentials=credentials)
+    except Exception as e:
+        print(f"Authentication failed: {e}")
+        print("Please ensure you have application default credentials set up.")
+        return
+
+    sql_map = get_sql_to_sheet_map(sql_dir)
+    response = sheets_service.spreadsheets().get(spreadsheetId=spreadsheet_id, includeGridData=False).execute()
+    sheets = response.get('sheets', [])
+
+    for sheet in sheets:
+        sheet_name = sheet['properties']['title']
+        sheet_id = sheet['properties']['sheetId']
+        charts = sheet.get('charts', [])
+
+        sql_file = sql_map.get(sheet_name)
+        if not sql_file:
+            # Try to match case-insensitively or show warning
+            sql_file = "TODO.sql"
+
+        for chart in charts:
+            title = chart['spec'].get('title', 'Untitled Chart')
+            chart_id = chart['chartId']
+
+            # Slugify for image name
+            image_name = re.sub(r'[^a-z0-9]+', '-', title.lower()).strip('-') + ".png"
+
+            # Construct markup
+            markup = f"""{{{{ figure_markup(
+image="{image_name}",
+caption="{title}",
+description="",
+chart_url="https://docs.google.com/spreadsheets/d/e/{PUBCHART_ID}/pubchart?oid={chart_id}&format=interactive",
+sheets_gid="{sheet_id}",
+sql_file="{sql_file}"
+)
+}}}}"""
+            print(markup)
+            print()
+
+
+if __name__ == "__main__":
+    # Resolve relative SQL_DIR based on script location
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+    absolute_sql_dir = os.path.normpath(os.path.join(script_dir, SQL_DIR))
+
+    print(f"Processing Spreadsheet: {SPREADSHEET_ID}")
+    print(f"SQL Directory: {absolute_sql_dir}\n")
+
+    generate_figure_markup(SPREADSHEET_ID, absolute_sql_dir)
diff --git a/sql/util/requirements.txt b/sql/util/requirements.txt
@@ -1,5 +1,7 @@
 pandas==2.3.3
 google-cloud-bigquery==3.40.0
+google-api-python-client==2.188.0
+google-auth==2.47.0
 requests==2.32.5
 tabulate==0.9.0
 gspread==6.2.1

diff --git a/src/.gitignore b/src/.gitignore
@@ -9,3 +9,4 @@ templates/sitemap.xml
 static/html/
 static/js/web-vitals.js
 .coverage
+eng.traineddata