Let me first carefully process the data.

Raw CSV rows (order_id, order_date, region, product, amount_usd):

1. 1001, 2025-01-06, Texas, widget, "1,240.00" → amount 1240.00, region Texas, date January → Q1
2. 1002, 1/9/2025, CA, gadget, $980 → 980, California, January → Q1
3. 1003, 2025-01-14, TX, gizmo, 312.50 → 312.50, Texas, Q1
4. 1004, Jan 21, 2025, california, widget, 1,764.00 → California Q1
5. 1005, 2025-01-27, Ohio, gadget, 455 → Ohio Q1
6. 1006, 2/3/2025, "texas " (trailing space), widget, N/A → excluded (not numeric), Texas
7. 1007, 2025-02-10, OH, gizmo, 1,102.25 → Ohio Q1
8. 1008, Feb 14, 2025, California, widget, 2,310.00 → California Q1
9. 1009, 2025-02-18, TX, gadget, -450.00 → Texas refund Q1
10. 1010, 2025-02-24, Texas, gizmo, 870.10 → Texas Q1
11. 1011, 3/3/2025, Ohio, widget, $1,999.99 → Ohio Q1
12. 1012, 2025-03-08, CA, gizmo, 640.00 → California Q1
13. 1008 (duplicate of 1008), Feb 14, 2025, California, widget, 2,310.00 → duplicate, remove
14. 1013, Mar 12, 2025, tx, gadget, 1,580.00 → Texas Q1
15. 1014, 2025-03-19, ohio, gizmo, 225.75 → Ohio Q1
16. 1015, 2025-03-27, California, gadget, 3,050.00 → California Q1
17. 1016, 4/2/2025, TX, widget, 990.00 → Texas Q2
18. 1017, 2025-04-07, Ohio, gadget, unknown → exclude, Ohio
19. 1018, Apr 11, 2025, CA, widget, 1,420.50 → California Q2
20. 1019, 2025-04-16, Texas, gizmo, 2,075.00 → Texas Q2
21. 1020, 2025-04-22, OH, widget, $760.25 → Ohio Q2
22. 1021, 5/1/2025, California, gizmo, 1,111.11 → California Q2
23. 1022, 2025-05-06, TX, gadget, 845.00 → Texas Q2
24. 1023, May 13, 2025, Ohio, gizmo, -220.00 → Ohio Q2 refund
25. 1024, 2025-05-19, texas, widget, 1,300.00 → Texas Q2
26. 1025, 2025-05-28, CA, gadget, 2,940.40 → California Q2
27. 1026, 6/4/2025, Ohio, widget, 515.00 → Ohio Q2
28. 1027, 2025-06-10, California, widget, 1,875.25 → California Q2
29. 1028, Jun 17, 2025, TX, gizmo, N/A → exclude, Texas
30. 1029, 2025-06-23, OH, gadget, 1,660.00 → Ohio Q2
31. 1030, 2025-06-30, Texas, gadget, 730.90 → Texas Q2

Total rows received: 31 (the number of data rows). Duplicates removed: 1 (the second occurrence of 1008). Rows excluded (non-numeric amounts): 1006 (N/A), 1017 (unknown), 1028 (N/A) → 3 excluded. Rows counted: 31 − 1 − 3 = 27.

Wait — are excluded rows counted from all received rows, or after deduplication? None of the excluded ones are duplicates, so counted = 31 − 1 − 3 = 27. Good.

Now let me compute the totals by region × quarter.

Texas Q1:
- 1001: 1240.00
- 1003: 312.50
- 1009: -450.00
- 1010: 870.10
- 1013: 1580.00
Sum: 1240 + 312.50 = 1552.50; −450 → 1102.50; +870.10 → 1972.60; +1580 → 3552.60.
Texas Q1 = 3552.60

Texas Q2:
- 1016: 990.00
- 1019: 2075.00
- 1022: 845.00
- 1024: 1300.00
- 1030: 730.90
Sum: 990 + 2075 = 3065; +845 = 3910; +1300 = 5210; +730.90 = 5940.90.
Texas Q2 = 5940.90
Texas total = 3552.60 + 5940.90 = 9493.50

California Q1:
- 1002: 980
- 1004: 1764.00
- 1008: 2310.00
- 1012: 640.00
- 1015: 3050.00
Sum: 980 + 1764 = 2744; +2310 = 5054; +640 = 5694; +3050 = 8744.
California Q1 = 8744.00

California Q2:
- 1018: 1420.50
- 1021: 1111.11
- 1025: 2940.40
- 1027: 1875.25
Sum: 1420.50 + 1111.11 = 2531.61; +2940.40 = 5472.01; +1875.25 = 7347.26.
California Q2 = 7347.26
California total = 8744 + 7347.26 = 16091.26

Ohio Q1:
- 1005: 455
- 1007: 1102.25
- 1011: 1999.99
- 1014: 225.75
Sum: 455 + 1102.25 = 1557.25; +1999.99 = 3557.24; +225.75 = 3782.99.
Ohio Q1 = 3782.99

Ohio Q2:
- 1020: 760.25
- 1023: -220.00
- 1026: 515.00
- 1029: 1660.00
Sum: 760.25 − 220 = 540.25; +515 = 1055.25; +1660 = 2715.25.
Ohio Q2 = 2715.25
Ohio total = 3782.99 + 2715.25 = 6498.24

Column totals:
Q1: 3552.60 + 8744.00 + 3782.99 = 16079.59
Q2: 5940.90 + 7347.26 + 2715.25 = 16003.41
Grand total: 16079.59 + 16003.41 = 32083.00
Total check: 9493.50 + 16091.26 + 6498.24 = 32083.00 ✓

Top region: California, $16,091.26.

Now, should I compute all of this from raw data in JavaScript in the HTML? The requirement says "the displayed numbers must be correctly computed from the data." Best approach: embed the raw CSV into the page, have JS do the cleaning and aggregation, and render the dashboard. That's more convincing and self-contained.

Plan: store CSV as template literal. Parse with a small CSV parser that handles quoted fields (containing commas). Then:

- Deduplicate on order_id (first occurrence).
- Normalize regions: build canonical map:
  - texas, TX, tx, Texas → "Texas" (with trailing space trim applied)
  - CA, california, California → "California"
  - OH, ohio, Ohio → "Ohio"
  Normalize: trim, lowercase, map aliases.
- Parse amounts: strip $ and commas; if parseFloat result is NaN → excluded.
- Parse dates: formats:
  - ISO yyyy-mm-dd
  - m/d/yyyy
  - "Mon d, yyyy" with month name abbreviation
  Get month number → quarter.
- Aggregate revenue[region][quarter].

Compute quality stats: totalRows, dupDropped, excluded, counted.

Render:
- Heading: "California is the top region at $16,091.26" — but compute dynamically.
- Pivot table with row/column totals.
- SVG bar chart of revenue by region. SVG generated via JS, scaled to max value. Horizontal or vertical bars. Vertical bars with labels.
- Quality panel.

Formatting: dollars with 2 decimals via toLocaleString or custom fmt.

Edge cases: floats — sums like 8744 might become 8744.000000000001? With 2-decimal money values, JS float error is negligible at this scale; use Math.round(x*100)/100 on output, and format with toFixed(2) anyway.

CSV parser: handle quoted fields with commas. Write simple parser: split lines, then char-by-char parse each line respecting quotes.

Let me write the parser:

```js
function parseCSV(text){
  const rows=[]; let row=[]; let field=''; let inQ=false;
  for(let i=0;i<text.length;i++){
    const c=text[i];
    if(inQ){
      if(c==='"'){ if(text[i+1]==='"'){field+='"';i++;} else inQ=false; }
      else field+=c;
    } else {
      if(c==='"') inQ=true;
      else if(c===','){row.push(field);field='';}
      else if(c==='\n'){row.push(field);field='';rows.push(row);row=[];}
      else if(c==='\r'){ /* skip */ }
      else field+=c;
    }
  }
  if(field.length||row.length){row.push(field);rows.push(row);}
  return rows;
}
```

Note: \r handling — if I skip \r when not in quotes, then "field" won't include \r. But "Jan 21, 2025" has quotes so inside quotes the \r would be preserved… not present in the file (CSV has \n only). But for safety: in quotes, only treat " specially; otherwise append — \r in quotes would get appended. That's fine; input is a template string I control. Also trailing content on final row.

Handle \r when not inQ: else if(c==='\r' && text[i+1]==='\n') continue; … skip. I control the raw data embedded, so I can use \n only. Keep parser simple: if c==='\r' skip.

Date parsing:
```js
function parseDate(s){
  s=s.trim();
  // ISO
  let m=/^(\d{4})-(\d{1,2})-(\d{1,2})$/.exec(s);
  if(m) return {m: +m[2]};
  // m/d/yyyy
  m=/^(\d{1,2})\/(\d{1,2})\/(\d{4})$/.exec(s);
  if(m) return {m:+m[1]};
  // "Mon name month day, year": e.g. Jan 21, 2025 — full or abbreviated month name
  m=/^([A-Za-z]{3,9})\s+(\d{1,2}),?\s+(\d{4})$/.exec(s);
  if(m){ const names={jan:1,feb:2,mar:3,apr:4,may:5,jun:6,jul:7,aug:8,sep:9,oct:10,nov:11,dec:12};
    return {m:names[m[1].slice(0,3).toLowerCase()]};
  }
  return null;
}
```
Wait, May 13 has "May" — month abbreviation fine.

Amount parsing:
```js
function parseAmount(s){
  s=s.trim();
  if(s==='') return NaN;
  const cleaned=s.replace(/[$,"]/g,''); // strip $ and commas
  if(isNaN(+cleaned)) return NaN;
  return +cleaned;
}
```
"unknown" → cleaned "unknown" → NaN. N/A → NaN. Good. Negative values preserved (replace doesn't strip '-'). Good.

Region normalization:
```js
const REGION_MAP={'tx':'Texas','texas':'Texas','ca':'California','california':'California','oh':'Ohio','ohio':'Ohio'};
function normRegion(s){ const key=s.trim().toLowerCase(); return REGION_MAP[key]||s.trim(); }
```
"texas " trailing space → trim → texas → Texas. "california" → California. Good.

Aggregation: regions sorted in fixed order? Sort regions alphabetically: California, Ohio, Texas. Quarter columns are Q1, Q2 (data covers only Q1/Q2 but compute dynamically: gather quarters). Quarter from month: q = Math.floor((month-1)/3)+1.

Pivot table: rows = regions (sorted), columns = quarters sorted numerically. Cell values, row totals, column totals, grand total. Formatting: fmt(v) = '$' + v.toLocaleString('en-US',{minimumFractionDigits:2,maximumFractionDigits:2}).

Bar chart: SVG. Compute max region total; bar heights proportional. Vertical bars: 3 regions, width ~600, labels below, value labels above. Handle negative totals (not in this case). Include axis baseline.

Quality panel: rows received = raw data row count (31), duplicates removed (1), excluded (3), counted (27).

Heading: top region by total revenue: compute max. "Top region: California — $16,091.26".

Also render a small note about excluded row ids? Optional, quality panel is fine as is.

Let me write the full HTML with nice styling. Add the table and maybe a chart. No external libraries, single file.

Let me draft the file.

```html
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<title>Sales Dashboard</title>
<style>
  body { font-family: ...; background:#f4f6f8; margin:0; padding:24px; color:#1e2430 }
  .card {...}
  table {...}
</style>
</head>
<body>
<h1>...
<div id="headline">
<div class="grid">
  <div class="card"><h2>Revenue by Region & Quarter</h2><div id="pivot"></div></div>
  <div class="card"><h2>Revenue by Region</h2><div id="chart"></div></div>
  <div class="card"><h2>Data Quality</h2><div id="quality"></div></div>
</div>
<script>
const RAW = `order_id,...`;
// processing
...
</script>
</body>
</html>
```

Chart: horizontal bars might be nicer. Vertical with labels also fine. I'll make horizontal bars: each region row, bar width proportional to max. SVG with text labels. Horizontal bars scale nicely with few categories.

Let me finalize the SVG: width 100%, viewBox "0 0 500 H". Each bar: height 28, gap. Left region name label, bar rect, value at right end of bar.

Chart also includes bar chart of "revenue by region" = totals. OK.

Compute totals dynamically.

Let me write JS:

```js
const rawRows = parseCSV(RAW.trim());
const header = rawRows.shift();
```
Wait, parser: rows include header. Good.

Process:
```js
const seen = new Set();
let dupDropped=0, excluded=0, counted=0;
const byRegionQ = {}; // region -> {q -> sum}
const rowsReceived = rawRows.length;
for(const r of rawRows){
  const [id, date, region, product, amount] = r;
  if(seen.has(id)){ dupDropped++; continue; }
  seen.add(id);
  const amt = parseAmount(amount);
  if(!isFinite(amt)){ excluded++; continue; }
  const d = parseDate(date);
  if(!d){ excluded++; continue; } // won't happen
  const q = Math.floor((d.m-1)/3)+1;
  const reg = normRegion(region);
  byRegionQ[reg] = byRegionQ[reg] || {q:{}};
  ...
  counted++;
}
```

counted = rowsReceived − dupDropped − excluded. Can also derive as counted = rowsReceived − dupDropped − excluded.

Pivot build: regions = Object.keys sorted; quarters = sorted distinct quarters.

Heading: top region.

Formatting function:
```js
const fmt = v => '$' + v.toLocaleString('en-US',{minimumFractionDigits:2, maximumFractionDigits:2});
```
Negative: toLocaleString gives "$-450.00"; acceptable? Negative cells won't appear in pivot (region-quant totals are all positive). But for robustness format negatives as -$450.00: handle sign.

fmt = v => (v<0?'-$':'$') + Math.abs(v).toLocaleString(...).

Bar chart: totals per region, sort descending.

Also verify computed values match my manual check. JS will compute so should be fine.

Rounding: totals might become 8744.000000000001? Sums like 980+1764+2310+640+3050 are integers exactly representable. 1420.50+1111.11+2940.40+1875.25: .50+.11+.40+.25 = 1.26 → might be 7347.259999... in floating point. toFixed(2) / toLocaleString with maximumFractionDigits:2 handles display rounding. To be safe, round all aggregates to 2 decimals: r2 = v=>Math.round(v*100)/100.

Apply r2 on final values.

Also check grand total consistency: pivot cells and totals computed from same underlying cell values → consistent.

Now write the final HTML. Include a table styling, card layout via CSS grid, headline style.

Also include excluded order ids as a note in quality panel? "Rows excluded: 3". That's fine.

Double-check row count: data rows 1001..1030 = 30 + duplicate 1008 = 31. Good.

Chart SVG: build via string. viewBox width 640. Bar per region row. Also draw max-value gridlines? Keep simple: bars with value label.

Let me write:

```js
const regionTotals = regions.map(reg => ({reg, total: regionsTotals...}));
```

Compute region total = sum over quarters. Sort desc for chart.

SVG:
```js
const maxV = Math.max(...totals.map(t=>t.total));
const barH=34, gap=18, padL=110, padR=70, W=640;
let svg = `<svg viewBox="0 0 ${W} ${...}" ...>`;
totals.forEach((t,i)=>{
  const y = i*(barH+gap)+10;
  const w = Math.max(2, (t.total/maxV) * (W - padL - padR));
  svg += `<text x="${padL-10}" y="${y+barH/2+5}" text-anchor="end" class="lbl">${t.reg}</text>`;
  svg += `<rect x="${padL}" y="${y}" width="${w}" height="${barH}" rx="4" fill="#3b82f6"/>`;
  svg += `<text x="${padL+w+8}" y="${y+barH/2+5}" class="val">${fmt(t.total)}</text>`;
});
```
Height = n*(barH+gap)+20. Negative total handling: clamp width, value label at left of bar? Skip complexity: assume non-negative (true for this data). Actually for robustness: if total<0 draw bar leftward? Overkill. w = max(2, ...) using Math.abs handles negatives. Keep simple but guard: if total<0, place label at padL + maxBarWidth. Good — minor.

Table styling with totals row/column bold.

Headline: `<strong>California</strong> leads all regions with $16,091.26 total revenue.`

Quality panel as a small table: Received: 31, Duplicates removed: 1, Excluded (bad amounts): 3, Counted: 27.

Also list excluded ids via tooltip? For insight include excluded order ids in the quality panel: collect excluded ids. Compute excludedIds list. Same for duplicate ids. Nice detail.

Now let me write out the whole file carefully.

parseDate regex for month name with comma: "Jan 21, 2025" — regex `^([A-Za-z]{3,9})\.?\s+(\d{1,2}),?\s*(\d{4})$`. My format: /^([A-Za-z]{3,9})\s+(\d{1,2}),?\s+(\d{4})$/ — "Jan 21, 2025" has space after comma; \s+ handles. "Feb 14, 2025" fine. "May 13, 2025" fine.

Month map via indexOf on 'janfeb...'? Use names array: ['jan',...] findIndex+1.

Amount parse: strip commas/$ from s, parseFloat, check isFinite. "1,240.00" → "1240.00" → 1240. "$1,999.99" → 1999.99. "-450.00" → -450. "N/A" → NaN via parseFloat → excluded. Good.

One caution: empty string parseFloat → 0? parseFloat('')→NaN. ok.

Also amount field parsed via CSV parser from raw: quoted "1,240.00" parses to `1,240.00` without quotes. Good.

Dedup: first occurrence kept: row index 7 (1008) kept, duplicate at index 12 dropped. seen set per raw data order. Good.

Order_id: string "1008" both. Good.

Region map: keys lowercased. "CA" → ca → California. "california" → California. "TX" → tx → Texas. "ohio" → ohio → Ohio. "texas " trim → texas. Good.

Quarter columns: collect set from data → [1,2]. Header labels "Q1","Q2".

Table header: Region | Q1 | Q2 | Total; then rows; then totals row "Total"; grand-total cell.

Now let me format the code. Use template literals in the page to build pivot HTML.

Also verify grand total: 32083.00 → fmt "$32,083.00".

OK let me write it all. Ensure RAW is exact CSV content provided.

Embedded CSV uses template literal backticks; content has no backticks. Double quotes fine. `${` not present. Good.

Careful: CSV contains quotes around fields; template literal preserves them.

Now finalize the HTML.

Also title header maybe: "Sales Dashboard — 2025 Q1–Q2". Fine.

Final answer: just the HTML file.