NLPify GCC News Sentiment Dataset

**Overview**
NLPify GCC News Sentiment Dataset provides sentiment scores on GCC equities sourced from multiple financial news APIs that aggregate Middle Eastern financial news from major news outlets and publications.

Sentiment scores and related metadata are published on a daily basis at 4pm Dubai time.

**Dataset type:** News
**Sentiment Feed Type:** Structured
**Applicable Asset Classes**: Publicly Listed Companies

**Markets Covered:**

1. Saudi Exchange (Tadawul)
2. Abu Dhabi Exchange (ADX)
3. Dubai Financial Markets (DFM)
4. Nasdaq Dubai
5. Qatar Stock Exchange (QSE)
6. Boursa Kuwait
7. Bahrain Bourse
8. Muscat Stock Exchange (MSX)

**Use cases**
Our client portfolio are currently using NLPify GCC News Sentiment Dataset in the following areas:

- Sales Trading
- Sell-Side Research
- Treasury
- Discretionary Asset Management
- Risk Management (via sentiment-driven surveillance)
- Hedge Funds
- Investor Relations

**Product details**
The *data schema* is organized in 11 distinct metrics:

- published_date
- ticker
- sentiment_daily
- sentiment_7MA
- sentiment_30MA
- sentiment_90MA
- attention_buzz_daily
- attention_buzz_7RSUM
- attention_buzz_30RSUM
- attention_buzz_90RSUM
- attention_buzz_weighted_sentiment

Description of *table fields*:
- PUBLISHED_DATE: The date and time at which the news headline was published, formatted as YYYY-MM-DD HH:MM:SS. This timestamp provides precise information on when the news item became public in GST (UTC+4) timezone.
- TICKER: The stock symbol for the company as per the trading system of the exchange it's listed on.
- SENTIMENT_DAILY: The sentiment score derived from the analysis of the news headline for the ticker on the specific published date and time, ranging from -1 (negative) to +1 (positive).
- SENTIMENT_7MA: The 7-day moving average of the news headline sentiment score for the ticker, smoothing out daily sentiment fluctuations to show a clearer trend.
- SENTIMENT_30MA: The 30-day moving average of the news headline sentiment score, providing a medium-term perspective on how sentiment from news headlines is trending for the company.
- SENTIMENT_90MA: The 90-day moving average of the news headline sentiment score, offering a long-term view of the company's sentiment trend based on news headlines.
- ATTENTION_BUZZ_DAILY: A measure of the media focus or interest in the ticker, quantified by the count of news headlines in which the ticker was mentioned.
- ATTENTION_BUZZ_7RSUM: The total media focus or interest in the ticker over the past 7 days, quantified by the count of news headlines in which the ticker was mentioned on a 7-day rolling sum basis.
- ATTENTION_BUZZ_30RSUM: The total media focus or interest in the ticker over the past 30 days, quantified by the count of news headlines in which the ticker was mentioned on a 30-day rolling sum basis.
- ATTENTION_BUZZ_90RSUM: The total media focus or interest in the ticker over the past 90 days, quantified by the count of news headlines in which the ticker was mentioned on a 90-day rolling sum basis.
- ATTENTION_BUZZ_WEIGHTED_SENTIMENT: The news headline sentiment score for the day, weighted by the volume of headlines the ticker received, combining sentiment with the level of attention.

**Business Needs**
- ***Sentiment Analysis***: Capture alpha-driving insights and uncover early risk signals with NLPify's cutting-edge Language AI and text quantification techniques. Our news sentiment dataset quantifies market breadth and asset sentiment, allows for further evidence to explain and interpret asset returns, and generates information edge which is not necessarily captured in "the numbers" or "the charts".
- ***Quantitative Analysis***: Predict equity movements and detect anomalies in real time by deploying NLPify's tickerized sentiment analysis. With our quantified textual sentiment scores and news attention buzz, you are able to construct alpha-generating quantitative trading programs and enhance existing legacy systematic strategies.
- ***Machine Learning***: Using NLPify's news sentiment scores and related metrics, you can train your own ML algorithm or use existing ML models to predict stock price and produce NLP-driven trading signals.
- ***Market Analysis***: Given that our news sentiment scores are mapped to individual stocks, sector and market analyses are facilitated by our dataset. To that extent, correlation analysis can be conducted among stocks from a sentiment vs. sentiment & sentiment vs. price perspectives. Further more, the dataset enables the discovery and evaluation of individual outliers within a portfolio, sector, or market. Also, it can play a key role in validating asset allocation strategies and portfolio rebalancing based on periodic sentiment shifts.
- ***Risk Analysis***: Risk management and capital preservation are enhanced substantially by using our news attention buzz-weighted sentiment scores. Spot risky assets based on sentiment-driven surveillance, detect unusual sentiment change in stocks, and minimize portfolio drawdowns by embedding sentiment analysis into your trading strategies. High negative sentiment accompanied by larger than usual news attention is an effective early risk indication of a critical company event that could precede a sudden price move.

For more details, refer to the embedded notebook.