Going pro in data science summary - AtmaMani/pyChakras Wiki
About the book
This is a short, easy to read book written in a conversational / blog style. The author shares his views on the industry and what it takes to succeed in data science and dispels some myth along the way. Get it from O'Rielly
- Skill sets most valuable are agile experimentation, hypothesis testing, professional data science programming.
- Finding useful evidence, interpreting its significance is the key skill of a practicing scientist, even more than mastering the machine learning algorithms.
- Before applying a tool, master the theoretical background. Some steps when applying a new tool
- find a problem
- choose a tool
- produce an output using the tool
- tinker with the tool until it addresses step 1.
On being successful
- Practice agile development. Work in small iterations, pivot based on results and learn along the way. Being agile does not guarantee an idea will succeed. But it decrease the time it takes to spot a dead end.
- Develop an active network to which you can share your work, get feedback and learn from others. Develop partners that are willing to experiment or implement recommendations from an analysis.
- Data science and big data in any org follows the typical hype cycle. There is a flood of enthusiasm and hope in the beginning. But as the project gets underway and first results are produced, the org plummets into disillusionment over the difference between what was imagined and what was obtained.
Overall, the author tries to convey that the skills required for a data scientist is similar to what you would expect from any other scientist working in any other domain. Skills such as a rooting in math, clearly understanding the problem, formulating a hypothesis and using the right tools to either falsify or prove the hypothesis are fundamental.