The current state of data science is sad. Our de facto model of data analysis is a waterfall, where we are unable to go back, after we have learned more about the things that we are trying to measure. An agile model for data science embraces our uncertainty about the real world, and allows us to iterate our analyses, hypotheses, and data collection. Clojure is uniquely suited for agile data science, as it provides a powerful set of core tools for data processing, and its flexibility and ecosystem allow you to build rapid prototypes of your own analysis tools. In this talk, I will demonstrate the strengths of Clojure ecosystem by a case example: a serverless Google Cloud pipeline for automatic processing and analysis of spoken interviews conducted by a chatbot. After preprocessing, the user of the system can explore the data using tailored analysis tools, which we were able to construct and iterate in just a few days.
Build your own tools for agile data science – Toni Vanhala |
| 2 Likes | 2 Dislikes |
| 412 views views | 1,162 followers |
| Science & Technology | Upload TimePublished on 17 Sep 2018 |
Không có nhận xét nào:
Đăng nhận xét