Which Data Scientist Type Are You?
There is an interesting Q&A about the two types of data scientists on Quora written by Michael Hochster, the director of research at Pandora.
Here’s a 30-second summary:
- Type A – A stands for analysis. The analytical data expert primarily uses statistics to mine, explore, and extract insights from data. She is comfortable with code, enough to support the analytical work.
- Type B – B stands for building. The type B data scientist is primarily a software engineer. She has some statistical background but is primarily best-suited for taking data insights and creating a product around it.
The dichotomy is a useful framework to think about radiology informatics as well.
It’s a wonderful read for anyone interested in data.
Radiologists Are Analytical by Nature
Radiologists generally lean heavily towards Type A. We are diagnostic experts who use knowledge of statistics to assess the veracity of research articles, and they are also able to use Bayesian reasoning, where appropriate, to inform a diagnostic process. Many radiologists have a coding background, but almost no one has enough experience to build commercial level products.
This probably explains why radiologists tend to have thoroughly thought-out ideas, often backed with insight, data, and creativity. However, because the typical radiologist-turned-informatician is a Type A data scientist, creating a tangible tool can be incredibly difficult.
Thus, perfect becomes the enemy of good.
Software engineers working for radiology vendors lean towards Type B. They have a reasonable understanding of radiology and the data necessary to produce good clinical work, but their primary contribution to the field arises from the ability to translate an idea into a product. The focus in this group is practicality, feasibility, and scalability.
Meaningful collaboration often requires both types of experts. If you have an interest in radiology informatics, knowing where you stand in this spectrum can help you determine where you can add most value. So, which data scientist type are you?