TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Show HN: I made an editable template to learn real-world Data Science

3 点作者 edylemond大约 3 年前

1 comment

edylemond大约 3 年前
This is an editable template to learn data science skills most companies are looking for (from my 3-year experience in London fintech as a data scientist, and from the self-learning journey leading up to that).<p>The study content provided in the template is minimal, but you can go as in-depth as you like with the linked resources. The idea is that you study those resources by yourself, and then write down what you learned in your own words, directly into your own copy of the template.<p>I like to learn with flashcards (especially to memorize common interview questions), so I’ve added some example flashcards to help you get started - you can add your own flashcards or delete them if it isn’t for you.<p>The template is based around 3 pillars:<p>- Math &amp; Stats<p>- Software Engineering &amp; Tools<p>- Data &amp; Business Communication<p>The Math &amp; Stats section contains a structured list of recommended topics and principles to learn, with links to relevant resources like Khan Academy videos and the classic books.<p>The Software Engineering &amp; Tools sections walks through tools to learn (based around the Jupyter-Python-Pandas ecosystem), and links to tutorials, videos, example notebooks and cheat sheets (all created by other fantastic people, I take no credit for the linked resources) to learn Python, Pandas, Scikit-Learn and Matplotlib.<p>The Data &amp; Business Communication section is the real core of the template, where both of the previous sections come together. It’s shaped after the process for a typical business data science project:<p>- Data collection<p>- Data exploration<p>- Data cleaning &amp; preparation<p>- Machine learning modeling: here I mention some common models actually used in businesses, like linear+logistic regression, random forests and timeseries forecasting<p>- Model evaluation<p>- Reporting &amp; data visualization: focus on creating clear plots here<p>- Communicating with stakeholders: this is where I go more in depth on communicating your results to business decision makers, and telling a story which a layman can understand