I\u2019ve released a research-grade financial dataset designed for machine<\/strong><\/p>\n learning and quantitative research, with a strong focus on preventing<\/strong><\/p>\n lookahead bias.<\/strong><\/p>\n The dataset includes:<\/strong><\/p>\n – Multi-asset daily price data<\/strong><\/p>\n – Technical indicators (momentum, volatility, trend, volume)<\/strong><\/p>\n – Macroeconomic features aligned by release dates<\/strong><\/p>\n – Risk metrics (drawdowns, VaR, beta, tail risk)<\/strong><\/p>\n – Strictly forward-looking targets at multiple horizons<\/strong><\/p>\n All features are computed using only information available at the time,<\/strong><\/p>\n and macro data is aligned using publication dates to ensure temporal<\/strong><\/p>\n integrity.<\/strong><\/p>\n The dataset follows a layered structure (raw \u2192 processed \u2192 aggregated),<\/strong><\/p>\n with full traceability and reproducible pipelines. A baseline,<\/strong><\/p>\n leakage-safe modeling notebook is included to demonstrate correct usage.<\/strong><\/p>\n The dataset is publicly available here:<\/strong><\/p>\n Kaggle link:<\/strong><\/p>\n https:\/\/www.kaggle.com\/datasets\/DIKKAT_LINKI_BURAYA_YAPISTIR<\/strong><\/a><\/p>\n Feedback and suggestions are very welcome.<\/strong><\/p>\n<\/div>\n submitted by \/u\/subcomandante_65 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
