https://rexarski.com/%E6%95%B0%E6%8D%AE%E7%A7%91%E5%AD%A6/Recent content in 数据科学 on howaboutqiuHugozh-CN© Qiū RuìWed, 26 Aug 2026 16:06:41 -0400https://rexarski.com/posts/2026/08/not-doing-data-science/Tue, 25 Aug 2026 22:30:24 -0400https://rexarski.com/posts/2026/08/not-doing-data-science/<p><a href="https://www.reddit.com/r/datascience/comments/1vx150x/im_a_data_scientist_but_i_dont_do_any_data_science/">reddit 上一则讨论</a>：发帖人说自己工作虽然是所谓的 data scientist 但是自己从来做的都不是真正的数据科学。原帖内容已经在一天后删除了。为什么删除呢？耐人寻味。</p> <p>票数最高的回答：</p> <blockquote> <p>Not so hot take, most companies problems are so large that you don’t need any data science to solve them. What you do need is a clean set of data, the right framing, and the ability to add and subtract. Nine times out of 10 that beats any fancy data science model.</p> <p>The impact of the insight is always more important than the method to derive the insight.</p>