Skip to content

openthings/databook-old

Repository files navigation

Databook-数据之书

Documentation Binder

我的剑传给能挥舞它的人。
       -- 查理.芒格

用于数据分析的Jupyter Notebooks。不需购买服务器,快速开始自己的数据分析过程。

创建步骤

  • 创建一个 github.com 账号。
  • 复制本代码仓库到自己的repository,点击fork即可。
  • 打开 https://mybinder.org , 填入上面的github repository地址,点击“launch”开始构建自己的远程服务。
  • 服务网页打开后,将后缀/tree改为/lab,以使用最新的多窗口界面。
  • 添加notebook, 然后运行、修改,shift+enter即可执行cell中的代码。
  • 提交更新到自己的 github repository, 通过运行 ./commit.sh。
  • https://github.com/openthings/databook 拉取更新, 运行 ./upstream-pull.sh.
  • 提交一个 pull-request, 将自己的 notebook 合并到上游 upstream, 需要从 github webUI上操作。

快速使用教程

安装的支持库

  • jupyter, 包含 notebook, jupyterhub, jupyterlab webUI.
  • conda, 功能强大的python科学数据分析版, 使用 miniconda更小,只包含基本安装包。
  • mlflow,使用 mlflow进行机器学习,该库由databricks公司进行开发。
  • airflow, 数据流水线处理工具-airflow, 由Apache Software Foundation资助开发。
  • tushare, 金融数据获取sdk,提供中国金融相关数据, 基于python api。
  • matplotlib, 图形绘制库,在 jupyter notebook 中使用,先执行 %matplotlib inline。
  • ipyleaflet , 在Jupyter Notebook中显示地图,提供jupyterlab extension, 是 jupyter-widgets的一部分。

更多资源