sohamphanseiitb/BIG_Data_5MSEC: BIG Data Analysis of NASA's 5 Millennium Solar Eclipse Database

Solar eclipses are a topic of interest among astronomers, astrologers and the general public as well. There were and will be about 11898 eclipses in the 5 millennia from 2000 BC to 3000 AD. Data visualization and regression techniques offer a deep insight into how various parameters of a solar eclipse are related to each other. Physical models can be verified and can be updated based on the insights gained from the analysis. The study covers the major aspects of data analysis including data cleaning, pre-processing, EDA, distribution fitting, regression and machine learning based data analytics. We provide a cleaned and usable database ready for EDA and statistical analysis.日食是天文学家、占星学者以及普通民众均关注的研究议题。在公元前2000年至公元3000年的五千年时间跨度内，共计存在约11898次日食事件（涵盖已发生的历史日食与未来将出现的日食）。数据可视化与回归分析技术可深入揭示日食各项参数间的内在关联规律。基于分析所得的研究结论，能够对物理模型进行验证与迭代更新。本研究覆盖了数据分析的核心流程，包括数据清洗、数据预处理、探索性数据分析（Exploratory Data Analysis, EDA）、分布拟合、回归分析以及基于机器学习的数据分析方法。我们已构建完成经过清洗的可用数据库，可直接用于探索性数据分析与统计分析工作。