Original dataset (has problems and needs cleaning before you use it -- this is why I give you a cleaned one)

Mohammed-Khalil Ghali, World Bank, "Commodities prices with yearly news (1960-2023) dataset", IEEE Dataport, August 22, 2025, doi:10.21227/rm91-bj66

https://ieee-dataport.org/documents/commodities-prices-yearly-news-1960-2023-dataset

This dataset provides a comprehensive annual time series of global commodity prices from 1960 to 2023, covering 31 different commodities across major categories including energy (crude oil, natural gas), agricultural products (cocoa, coffee, tea, various oils, grains), food items (rice, wheat, fruits), livestock products (beef, chicken), industrial materials (cotton, rubber), fertilizers (phosphate rock, potassium chloride), and metals (aluminum, iron ore, copper, precious metals like gold, platinum, and silver). Each year's entry includes the average prices for these commodities along with a "Commodity News Headline" field containing relevant market developments and a list of the top most significant commodities news headlines for that year, alongside their corresponding embeddings conversion making this dataset valuable for analyzing long-term commodity price trends, market volatility, and the historical relationships between different sectors of the global commodity markets over more than six decades.