charlieviettq/awesome-agent-skill / .claude/skills/algo-social-sentiment
"algo-social-sentiment"
Verified"Implement VADER sentiment analysis for social media text scoring. Use this skill when the user needs to analyze sentiment in tweets, reviews, or social posts, compute compound sentiment scores, or classify text polarity — even if they say 'is this positive or negative', 'sentiment of these comments', or 'social media mood analysis'.".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.9 KB |
| examples/sample_scenario.md | markdown | 4.6 KB |
| references/model-comparison.md | markdown | 9.5 KB |
| references/vader-rules.md | markdown | 10.3 KB |
examples/sample_scenario.md
View on GitHub# Example: 電商品牌新品上市後的社群情緒監測
## Scenario
**背景:** TrendKit(台灣中型運動用品電商)於 2026-03-15 推出新款無線耳機「TrendKit AirFlow Pro」,上市後一週內收到大量社群反應。行銷團隊從 Twitter/X 和 Dcard 爬取了 450 則留言,想在週一晨會前量化情緒分佈,判斷是否需要緊急公關應對。
**使用者問題:** 「這些留言整體是正面還是負面?有沒有哪個議題特別拉低評分?」
**原始資料樣本(450 則中的代表性摘錄):**
```
[1] "AirFlow Pro 音質真的超棒!戴一整天耳朵不痛 👍👍"
[2] "連線一直斷,退貨!!!"
[3] "還可以啦,沒特別驚艷"
[4] "AMAZING sound quality for the price 😍"
[5] "Oh great, another earphone that dies in 3 hours 🙄"
[6] "不難用但不值這個價"
[7] "降噪功能完全沒用 垃圾"
[8] "配戴感超舒服!通勤必備"
[9] "充電盒做工太差,塑膠感很重"
[10] "性價比算高啦,推薦給預算有限的人"
```
---
## Analysis
### Phase 1: Input Validation
- 共 450 則文字,混合繁體中文與英文。
- **注意:VADER 為英文設計**。中文留言需前處理:將常見中文情緒詞翻譯或標記後再送入 VADER,或分流處理(中文用 SnowNLP,英文用 VADER)。
- 本案採**分流策略**:英文或含英文關鍵字的留言走 VADER;純中文留言走 SnowNLP,最後合併結果。
- 過濾空白留言:0 則。Gate ✓
---
### Phase 2: Core Algorithm(VADER 對英文/混合留言)
以上 10 則為例,標記語言後處理:
| # | 語言判定 | 工具 | compound | label |
|---|---------|------|----------|-------|
| 1 | 中文為主 | SnowNLP | +0.72 | positive |
| 2 | 中文 | SnowNLP | −0.61 | negative |
| 3 | 中文 | SnowNLP | +0.08 | neutral |
| 4 | 英文 | VADER | +0.87 | positive |
| 5 | 英文 | VADER | +0.34 | **positive** ⚠️ |
| 6 | 中文 | SnowNLP | −0.22 | negative |
| 7 | 中文 | SnowNLP | −0.78 | negative |
| 8 | 中文 | SnowNLP | +0.65 | positive |
| 9 | 中文 | SnowNLP | −0.31 | negative |
| 10 | 中文 | SnowNLP | +0.41 | positive |
**留言 [5] Sarcasm 陷阱:**
> "Oh great, another earphone that dies in 3 hours 🙄"
VADER 看到 "great" → compound = +0.34,判為正面。但語意明顯負面(諷刺)。VADER 無法偵測反諷,此類留言需人工抽樣複查。
---
### Phase 3: Verification(全量 450 則)
**人工 spot-check(隨機抽 20 則):**
- 18 則分類與人工判斷一致
- 2 則誤判(均為英文諷刺句,結構為 "great/perfect + 負面結果")
- 準確率 90%,超過 80% 門檻 → Gate ✓
**議題分群(關鍵詞 TF-IDF 後人工標記):**
| 議題 | 留言數 | 平均 compound |
|------|--------|--------------|
| 音質 / 降噪 | 112 | +0.58 |
| 配戴舒適度 | 89 | +0.61 |
| 藍牙連線穩定 | 74 | **−0.52** |
| 電池續航 | 68 | **−0.41** |
| 充電盒做工 | 43 | −0.28 |
| 性價比 | 64 | +0.33 |
---
### Phase 4: Output
```json
{
"results_sample": [
{"text": "AMAZING sound quality for the price 😍", "compound": 0.87, "pos": 0.65, "neu": 0.35, "neg": 0.0, "label": "positive"},
{"text": "Oh great, another earphone that dies in 3 hours 🙄", "compound": 0.34, "pos": 0.28, "neu": 0.72, "neg": 0.0, "label": "positive"},
{"text": "連線一直斷,退貨!!!", "compound": -0.61, "pos": 0.0, "neu": 0.22, "neg": 0.78, "label": "negative"}
],
"metadata": {
"texts_analyzed": 450,
"date_range": "2026-03-15 to 2026-03-21",
"distribution": {
"positive": 0.48,
"neutral": 0.22,
"negative": 0.30
},
"overall_compound_mean": 0.14,
"sarcasm_suspected": 11,
"language_split": {"chinese_snowNLP": 318, "english_vader": 132}
}
}
```
---
## Result
**整體情緒:偏正面,但有明確痛點**
- 48% 正面、22% 中立、30% 負面;整體 compound 均值 +0.14,屬於「溫和正面」。
- **音質與配戴舒適**是口碑亮點(compound 均值 > +0.58)。
- **藍牙連線穩定性**是最大拉力(compound −0.52),74 則負評集中於「斷線」、「延遲」,建議工程團隊優先查韌體版本。
- **電池續航**為次要痛點(compound −0.41),68 則提及「3 小時就沒電」,與官方宣稱 8 小時出入明顯,需公關釐清。
- 發現 11 則疑似諷刺留言被誤判為正面,已標記供公關人員人工複查。
**建議行動:**
1. 本週發布韌體更新說明,主動回應藍牙斷線問題。
2. 公關稿澄清電池續航測試條件(音量、降噪開啟與否)。
3. 每日跑一次情緒監測,追蹤韌體更新後負評比例是否下降。
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |