Skip to content
  • datasets
  • sentiment

News Sentiment Dataset

Access AI-powered sentiment analysis derived from financial news and social media. FinBrain’s NLP models process thousands of articles daily to generate sentiment scores that help you gauge market mood and momentum.

What’s Included

Section titled “What’s Included”

The News Sentiment dataset provides:

  • Sentiment Score: Normalized score from -1 (bearish) to +1 (bullish) for each date
  • Date-Keyed Data: Historical sentiment organized by date
  • Daily Updates: Fresh sentiment scores every trading day
  • Historical Data: Years of sentiment history for backtesting

Coverage

Section titled “Coverage”
MarketTickersUpdate Frequency
S&P 500500+Daily
NASDAQ3,000+Daily
NYSE2,500+Daily
Crypto100+Daily

Understanding Sentiment Scores

Section titled “Understanding Sentiment Scores”
RangeInterpretation
0.5 to 1.0Strong bullish sentiment
0.2 to 0.5Moderate bullish sentiment
-0.2 to 0.2Neutral sentiment
-0.5 to -0.2Moderate bearish sentiment
-1.0 to -0.5Strong bearish sentiment

Note: Sentiment scores are returned as strings and need to be converted to floats for numerical operations (e.g., float(score)).

Quick Start

Section titled “Quick Start”
from finbrain import FinBrainClient


fb = FinBrainClient(api_key="YOUR_API_KEY")


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

For complete code examples in Python, JavaScript, C++, Rust, and cURL, see the API Reference.

Use Cases

Section titled “Use Cases”

Sentiment-Based Trading Signals

Section titled “Sentiment-Based Trading Signals”

Generate trading signals based on sentiment thresholds:

from finbrain import FinBrainClient


fb = FinBrainClient(api_key="YOUR_API_KEY")


def get_sentiment_signal(market, ticker):
    """Generate trading signal from sentiment score"""
    sentiment = fb.sentiments.ticker(market, ticker)
    sentiment_data = sentiment["sentimentAnalysis"]


    if not sentiment_data:
        return "no_data"


    # Get the latest date's score (convert string to float)
    dates = sorted(sentiment_data.keys(), reverse=True)
    latest_score = float(sentiment_data[dates[0]])


    if latest_score > 0.5:
        return "strong_buy"
    elif latest_score > 0.2:
        return "buy"
    elif latest_score < -0.5:
        return "strong_sell"
    elif latest_score < -0.2:
        return "sell"
    else:
        return "hold"


signal = get_sentiment_signal("S&P 500", "TSLA")
print(f"Signal: {signal}")

Sentiment Screening

Section titled “Sentiment Screening”

Screen a watchlist for tickers with extreme sentiment:

from finbrain import FinBrainClient


fb = FinBrainClient(api_key="YOUR_API_KEY")


watchlist = ["AAPL", "GOOGL", "MSFT", "AMZN", "NVDA", "TSLA"]


bullish = []
bearish = []


for ticker in watchlist:
    sentiment = fb.sentiments.ticker("S&P 500", ticker)
    sentiment_data = sentiment["sentimentAnalysis"]
    if sentiment_data:
        dates = sorted(sentiment_data.keys(), reverse=True)
        score = float(sentiment_data[dates[0]])
        if score > 0.5:
            bullish.append((ticker, score))
        elif score < -0.5:
            bearish.append((ticker, score))


print("Bullish tickers:", bullish)
print("Bearish tickers:", bearish)

Combine with Price Predictions

Section titled “Combine with Price Predictions”

Enhance prediction confidence when sentiment aligns with expected price movement:

from finbrain import FinBrainClient


fb = FinBrainClient(api_key="YOUR_API_KEY")


def analyze_ticker(market, ticker):
    """High conviction signals when predictions and sentiment align"""
    predictions = fb.predictions.ticker(ticker, prediction_type="daily")
    sentiment = fb.sentiments.ticker(market, ticker)


    pred = predictions["prediction"]
    expected_short = float(pred["expectedShort"])


    sentiment_data = sentiment["sentimentAnalysis"]
    dates = sorted(sentiment_data.keys(), reverse=True)
    sent_score = float(sentiment_data[dates[0]])


    # Stronger signal when expected move and sentiment align
    if expected_short > 0.5 and sent_score > 0.3:
        return "high_conviction_buy"
    elif expected_short < -0.5 and sent_score < -0.3:
        return "high_conviction_sell"
    else:
        return "mixed_signals"


result = analyze_ticker("S&P 500", "AAPL")
print(result)
Section titled “Related Resources”