News sentiment analysis uses artificial intelligence to quantify the mood of financial news. Instead of reading hundreds of articles, traders can use sentiment scores to quickly gauge whether coverage is bullish, bearish, or neutral.

What is News Sentiment Analysis?

News sentiment analysis is the process of using natural language processing (NLP) and machine learning to extract the emotional tone from financial news articles, press releases, and other text sources.

The output is typically a sentiment score—a numerical value representing how positive or negative the coverage is for a particular stock or asset.

Typical sentiment scale:

+1.0 = Extremely bullish
+0.5 = Moderately positive
0.0 = Neutral
-0.5 = Moderately negative
-1.0 = Extremely bearish

How Sentiment Analysis Works

Modern sentiment analysis uses several techniques:

Technique	Description
Lexicon-based	Counts positive/negative words using financial dictionaries
Machine Learning	Models trained on labeled financial text
Deep Learning	Transformer models (BERT, GPT) that understand context
Hybrid	Combines multiple approaches for accuracy

Why Context Matters

Financial sentiment is nuanced. Consider these examples:

“Revenue fell 5%” → Negative
“Revenue fell only 5%, beating expectations” → Positive
“The company killed it this quarter” → Positive (idiom)

Advanced models understand these contextual differences that simple keyword counting would miss.

Why Sentiment Matters for Trading

News sentiment can be a leading indicator:

Use Case	Description
Early signal	Sentiment shifts before price moves
Confirmation	Validates technical or fundamental analysis
Risk management	Detect negative sentiment before position sizing
Contrarian plays	Extreme sentiment can signal reversals

Sentiment as a Contrarian Indicator

Extreme readings often precede reversals:

Sentiment Level	Potential Interpretation
Above +0.7	Excessive optimism—potential top
Below -0.7	Excessive pessimism—potential bottom
-0.3 to +0.3	Neutral—no strong directional bias

This works because markets tend to overshoot in both directions based on news cycles.

Accessing Sentiment Data

You can access historical sentiment scores through the FinBrain API:

from finbrain import FinBrainClient

fb = FinBrainClient(api_key="YOUR_API_KEY")

# Get sentiment data as DataFrame
df = fb.sentiments.ticker("S&P 500", "AAPL", as_dataframe=True)

print(df.tail())
#             sentiment
# date
# 2024-11-04      0.186
# 2024-11-01      0.339
# 2024-10-31      0.565

# Calculate sentiment statistics
print(f"\nMean sentiment: {df['sentiment'].mean():.3f}")
print(f"Current sentiment: {df['sentiment'].iloc[-1]:.3f}")

# Add moving averages for trend analysis
df["ma5"] = df["sentiment"].rolling(5).mean()
df["ma20"] = df["sentiment"].rolling(20).mean()

# Detect sentiment momentum
df["momentum"] = df["sentiment"] - df["ma5"]

# Find extreme readings
extreme_bullish = df[df["sentiment"] > 0.7]
extreme_bearish = df[df["sentiment"] < -0.7]

print(f"\nDays with extreme bullish sentiment: {len(extreme_bullish)}")
print(f"Days with extreme bearish sentiment: {len(extreme_bearish)}")

Building Sentiment Signals

Momentum Strategy

Trade in the direction of sentiment momentum:

def sentiment_momentum_signal(df):
    """Generate signal based on sentiment momentum"""
    current = df["sentiment"].iloc[-1]
    ma5 = df["sentiment"].tail(5).mean()
    ma20 = df["sentiment"].tail(20).mean()

    # Momentum crossover
    if ma5 > ma20 and current > 0.2:
        return "bullish"
    elif ma5 < ma20 and current < -0.2:
        return "bearish"
    else:
        return "neutral"

Contrarian Strategy

Trade against extreme sentiment:

def contrarian_signal(df, lookback=20):
    """Generate contrarian signal from extreme sentiment"""
    recent = df.tail(lookback)
    current = recent["sentiment"].iloc[-1]
    mean = recent["sentiment"].mean()
    std = recent["sentiment"].std()

    z_score = (current - mean) / std if std > 0 else 0

    if z_score > 2:
        return "contrarian_sell"  # Too optimistic
    elif z_score < -2:
        return "contrarian_buy"   # Too pessimistic
    else:
        return "neutral"

Combining Sentiment with Other Data

Sentiment works best when combined with other signals:

Sentiment + Price Action – Confirm breakouts with positive sentiment
Sentiment + Insider Buying – High conviction when both align
Sentiment + AI Predictions – Multiple models agreeing increases confidence
Sentiment + Technical Levels – Use sentiment to validate support/resistance

Key Takeaways

News sentiment analysis quantifies the mood of financial news
Scores typically range from -1 (bearish) to +1 (bullish)
Extreme readings can serve as contrarian indicators
Sentiment momentum can signal directional trades
Best used in combination with other data sources

Ready to add sentiment analysis to your trading strategy? Explore the News Sentiment Dataset and API Reference.

What is News Sentiment Analysis? A Guide for Traders