Market Sentiment → Brand Positioning

Scenario

Alpha Vantage’s NEWS_SENTIMENT function returns sentiment scores for tickers and topics across financial news. This job pulls sentiment data for your company’s ticker and competitors, maps the sentiment landscape, then asks Mave to recommend brand positioning adjustments — should you lean into market optimism, differentiate against pessimism, or stay neutral?

Flow: Alpha Vantage GET ?function=NEWS_SENTIMENT&tickers={ticker} → Mavera POST /mave/chat → Positioning recommendations

Code

Python

import os, requests, time

AV_KEY = os.environ["ALPHA_VANTAGE_KEY"]
AV_BASE = "https://www.alphavantage.co/query"
MV = os.environ["MAVERA_API_KEY"]
MV_BASE = "https://app.mavera.io/api/v1"
MV_H = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}

TICKERS = {"AAPL": "Apple", "MSFT": "Microsoft", "GOOGL": "Google"}
OUR_TICKER = "AAPL"

# 1. Fetch news sentiment per ticker
sentiment_data = {}
for ticker, name in TICKERS.items():
    r = requests.get(AV_BASE, params={
        "function": "NEWS_SENTIMENT", "tickers": ticker,
        "limit": 20, "apikey": AV_KEY,
    })
    r.raise_for_status()
    feed = r.json().get("feed", [])
    
    scores = []
    articles_summary = []
    for item in feed:
        for ts in item.get("ticker_sentiment", []):
            if ts.get("ticker") == ticker:
                score = float(ts.get("ticker_sentiment_score", 0))
                scores.append(score)
                articles_summary.append(
                    f"- {item.get('title','')[:100]} (score: {score:.2f}, "
                    f"relevance: {ts.get('relevance_score','N/A')})"
                )
    
    avg = sum(scores) / len(scores) if scores else 0
    sentiment_data[ticker] = {
        "name": name, "avg_sentiment": avg, "article_count": len(feed),
        "positive": sum(1 for s in scores if s > 0.15),
        "negative": sum(1 for s in scores if s < -0.15),
        "neutral": sum(1 for s in scores if -0.15 <= s <= 0.15),
        "articles": articles_summary[:10],
    }
    print(f"{name} ({ticker}): avg={avg:.3f}, articles={len(feed)}, "
          f"+{sentiment_data[ticker]['positive']}/-{sentiment_data[ticker]['negative']}")
    time.sleep(15)

# 2. Build sentiment landscape
landscape = "\n\n".join(
    f"## {d['name']} ({t})\n"
    f"Average Sentiment: {d['avg_sentiment']:.3f}\n"
    f"Distribution: {d['positive']} positive, {d['neutral']} neutral, {d['negative']} negative\n"
    f"Top Headlines:\n" + "\n".join(d['articles'][:5])
    for t, d in sentiment_data.items()
)

# 3. Mave positioning analysis
analysis = requests.post(f"{MV_BASE}/mave/chat", headers=MV_H, json={
    "message": f"Brand strategist specializing in market-sentiment-aligned positioning.\n\n"
        f"SENTIMENT LANDSCAPE (our ticker: {OUR_TICKER}):\n\n{landscape}\n\n"
        "Analyze:\n"
        "1. **Sentiment Map** — Where does each player sit? Who's gaining/losing favor?\n"
        "2. **Our Position** — Given our sentiment vs competitors, what positioning plays are available?\n"
        "3. **Messaging Recommendations** — 3 specific messaging angles based on the sentiment gap\n"
        "4. **Timing** — Is sentiment trending up/down? Should we act now or wait?\n"
        "5. **Risk** — What happens if sentiment reverses?\n\n"
        "Be specific. Reference actual headline themes."
}).json()

print(f"\n{'='*60}\nSENTIMENT-BASED POSITIONING ANALYSIS\n{'='*60}")
print(analysis.get("content", "")[:2500])

Example Output

Apple (AAPL): avg=0.187, articles=20, +12/-3
Microsoft (MSFT): avg=0.241, articles=20, +15/-2
Google (GOOGL): avg=0.098, articles=18, +8/-5

SENTIMENT-BASED POSITIONING ANALYSIS
============================================================

1. SENTIMENT MAP:
   Microsoft leads (+0.241) — driven by Azure AI and Copilot coverage.
   Apple mid-range (+0.187) — Vision Pro mixed but Services revenue strong.
   Google weakest (+0.098) — antitrust coverage dragging sentiment down.

2. OUR POSITION (AAPL):
   Healthy positive but gap vs. MSFT. Services narrative is winning;
   hardware skepticism (Vision Pro) is a drag. Opportunity: lean into
   the privacy + ecosystem story that differentiates from both.

3. MESSAGING RECOMMENDATIONS:
   a) "Privacy is the product" — exploit Google's antitrust weakness
   b) "Ecosystem, not features" — differentiate from MSFT's feature-race
   c) "Services momentum" — ride the positive revenue narrative

Error Handling

Free tier limits

25 req/day. Each ticker = 1 call. 3 tickers = 3 calls. Add 15s delay between calls to stay under 5/min. Cache results for same-day re-runs.

Empty feed

Some tickers return empty feeds. Check r.json().get("Note") for rate-limit messages — Alpha Vantage returns 200 with a note instead of 429.

Sentiment scale

Scores range -1 to +1. Thresholds: above 0.35 = bullish, below -0.35 = bearish, between = neutral. Adjust thresholds for your industry.