The-Thomson-Reuters-NewsScope-Event-Indices

Notes and Implementation of - Managing real-time risks and returns: The Thomson Reuters NewsScope Event Indices

Purpose

Manage event risk, which is posed by unanticipated news that causes major market moves over short time interval.

The Framework for Real-time News Analytics

The core of the real-time news analysis engine relied on analyzing market behaviors as a result of news from a specific categories such as market, macroeconomic or corporate news.

In this notebook, I analyze log returns of several ETFs when there is a shift of sentimet measured by zscore over 60 day rolling data.

Event Index

I extracted the following indices of news - ['BANKRUPTCY', 'CENTRALBANK', 'DEBT', 'DEFLATION', 'DEREGULATION', 'EARNINGSREPORT', 'GOLDPRICE', 'HOUSING_PRICES', 'INFLATION', 'INTEREST_RATES', 'IPO', 'MIDDLECLASS', 'OILPRICE', 'STOCKMARKET', 'TRADE_DISPUTE']

Files

1. tones.csv contains sentiment data of of several indices that are analyzed through few thousands news articles since 2017. Each score of "tone" is measured by average sentiment of artiles where related keyworkds appeared in.
2. etf_data.csv contains stock data downloaded from Yahoo Finance.
3. Event_Analysis.ipynb is the notebook that attempts to analyze the shifts of distribution of log returns as a results of shift of zscore of the sentiment.