hugodevere / fomcanalysis-model Goto Github PK

FOMCAnalysis Model Overview

Youtube tutorial: https://youtu.be/HZPNm65Xg7E

The FOMCAnalysis class provides a comprehensive analysis of FOMC (Federal Open Market Committee) data, with various functionalities and methods. Here's a breakdown of the different components of the code:

Class Initialization:

restart_data (default value: False): This flag indicates whether to restart the data or not. Its use would depend on the other methods within this class.

start_date: This specifies the start date for the analysis. There is no default value, so it must be specified when creating an instance of the class.

price_data_files: This parameter is expected to hold the names or paths of files that contain price data. Again, there is no default value, so it must be provided.

price_base_path: This is the base path where the price data files are located. This also must be specified by the user.

headline_data_file: This parameter should hold the file name or path for the headline data file. This must be provided by the user.

master_file_path: This parameter is expected to hold the file name or path for the master file. It must be specified by the user.

Data Loading Methods:

The code imports necessary libraries such as pandas, datetime, tqdm, and warnings.

load_price_data: Loads price data from CSV files and returns a combined DataFrame.

load_headline_data: Loads headline data from a CSV file and returns a DataFrame.

save_results_to_csv: Saves the analysis results to a CSV file.

update_master_file: Updates the master file with new trades.

Sentiment Analysis, Filtering and Ranking Methods:

analyze_headlines_sentiment: Performs sentiment analysis on headlines by applying predefined positive and negative terms to calculate sentiment scores. (Hawk vs Dove Score Index and Sentiment Index)

analyze_headlines: Analyzes the headlines, calculating sentiment scores for each headline based on price data and sentiment analysis results. (Market reaction Index)

filter_fedspeak: Filters Fed speak headlines from a DataFrame based on the names of officials and types of releases.

calculate_pnl: Calculates the profit and loss for each headline based on price data.

rank_headlines: Ranks the headlines based on various factors, including price, volume, and sentiment scores.

Main Method:

main: The main method conducts the FOMC analysis. It loads price data, headline data, analyzes the headlines, saves the results, and updates the master file.

extract_analysis: This method extracts the analysis results from the main method. It filters the Fedspeak data, ranks the headlines, calculates profit and loss, and returns a dictionary containing the filtered DataFrame, all headlines, and recent headlines.

The model output is a dictionary with the following key-value pairs:

filtered_df: A dictionary containing individual dataframes of headline analysis for each Fed official and press release.

All_headlines: A list of all headlines ranked in order of significance.

Recent_headlines: A list of the most recent headlines, specified by the start_date parameter, ranked by importance.

Understanding FOMCAnalysis Measures:

Market Reaction: This measure assesses the response to each Fedspeak-related headline. It considers the total volume traded from the headline release time (t) to two minutes after (t+2) and calculates the price change from three minutes before the headline (t-3) to two minutes after (t+2). If multiple headlines are released simultaneously, only the first one is considered.

Market Significance: Importance is determined by the rolling average of the absolute market reaction to corresponding Fedspeak from a given Federal Reserve official.

Sentiment Index: The sentiment index utilizes the VADER (Valence Aware Dictionary and sEntiment Reasoner) score, a sentiment analysis tool that quantifies the positivity, neutrality, or negativity of text. VADER employs a sentiment score dictionary and lexical heuristics, accounting for factors such as intensifiers, punctuation, and capitalization. It excels at analyzing short texts like social media content.

Hawk-Dove Score: Based on Tadle's methodology, this score indicates the sentiment behind the Federal Open Market Committee's (FOMC) monetary policy as expressed in their speeches. A higher (Hawkish) score signifies a policy inclination towards higher interest rates to curb inflation, while a lower (Dovish) score suggests a preference for lower interest rates to boost economic growth. This score aids in predicting market trends based on expected economic policy. See: Tadle, R. C. (2022). FOMC minutes sentiments and their impact on financial markets. Journal of Economics and Business.