Qualities that make a great data scientist
- The first and foremost, in my opinion, is curiosity. The core of science is to understand how things work, or “being curious”. As the term “scientist” in “data scientist”, a great scientist always wonder and question “why/why not?”. For data science problems, data cleansing/preparation takes most of the time. This step means to discover data and find the insight before any further step can be implemented. (if we couldn’t see the data pattern, properly the algorithms neither could.) Discovery is usually associated with curiosity. Understanding how things work, in case of data science viewpoint, is to understand how data are patterned. The intense curiosity can help data scientist dig deeper the information and engineer key features. This usually has a greater impact on the model performance than the step of algorithms tuning does.
- Secondly, to completely understand “data” in “data science” requires a strong relevant technical background (e.g. math, coding …). Curiosity is insufficient if a data scientist does not know which efficient tools/methods should be used to interpret data (especially in data preparation) in various situations. The wrong choice of methods can yield insufficient insight, which leads to poor results at the early stage of solving problems.
- Last but not least, communication skills are essential. Communication can be in the form of metrics or data visualization/presentation so that everyone could understand. Not everyone in the team is data scientist or having a similar background. The work may not be in production if data scientists cannot persuade the other teammates, especially the non-technical ones. If a data scientist can simply explain his/her work for everyone, then he/she understands the business problem well.
References:. https://www.quora.com/What-characteristics-make-for-a-good-data-scientist . https://www.quora.com/What-qualities-separate-great-data-scientists-from-good-data-scientists
By: Linh Viet Nguyen