Introduction
While I was working recently I noticed the difference between data science projects that incorporate the humanistic element and those that don’t. The ones that incorporate the humanistic element tend to be much more impactful and useful to the end stakeholder. This is because the humanistic side of data science ensures that the data scientist is taking into account the people who are affected by this and always works in a sense that is ethical and truthful to them. The human side of data science encompasses “the ability to ask the right questions, make ethical decisions, communicate effectively, and adapt to change”(source).
Due to this, we are going to be focusing on the human side of data going forward because it really is the difference between a good and a great data scientist. Starting with asking the right questions. Asking questions is easy but asking the right questions can be very hard. This is what the article “How To Ask The Right Questions As A Data Scientist” by Admond Lee, a data science entrepreneur, focuses on(linked here).
Summary of Article
The article starts out by emphasizing the importance of asking the right questions. Not asking the right questions and skipping over this step can cause weeks of work to be lost out, as did happen to the author one time. To ask the right questions they should all get at the problem statement. The author then gets into the 4 stages of the problem statement: understand the problem that needs to be addressed or solved, assess the situation with respect to the problem, understand the potential risks and benefits associated with the project, and define success criteria(or metric) to assess the project. Starting off with the first stage: understand the problem that needs to be addressed or solved. In the real world stakeholders will come to you with a problem and it is up to you to frame their business problem into a data science problem through empathy. You should also try to learn domain knowledge from them and combine it with your technical knowledge to deliver a result that drives business value. Next the second stage: assess the situation with respect to the problem. This step focuses on understanding any constraints that you have due to your situation or the situation of your stakeholders. Something as simple as computing power could be a constraint that you might have to work around. The article focuses on the third stage of the process: understanding the potential risks and benefits associated with the project. The author notes that this step is optional depending on your problem but can still help you. This step is all about assessing any benefits and risks that are associated with your project which can help assess the validity of the problems statement defined earlier. After, the author moves onto the last stage: defining success criteria(or metrics) to assess the project. This stage is especially important according to the author because you don’t want a very important project that you finish but you don’t know if it was successful because you don’t have anything to measure that. It all comes down to what you want out of the project. The measure, the author says, should be measurable and not ambiguous.
My Take
Overall the article was a good one. It explained everything well and through the use of headers was easy to read. I would have preferred some more pictures to help enhance the visual appeal. I really liked however the way that things were explained and how the author listed the topic in steps to make it easy to follow and implement into one’s workflow. One recommendation I do have is for the last step to make your goals SMART goals(specific, measurable, attainable, relevant, time-based). This will help make sure that your goals are specific and not ambitious such as the author said they should be. Not only that but it will also help you refine your goals and therefore will help you understand the project even more contributing to a positive cycle and enhancing the use of the humanistic side of data science.
Conclusion
All in all, the article was a good one. I really enjoyed its use of steps to make the whole article very followable and the methods very implementable into your workflow. The article was easy to read and follow. However, I do recommend using the SMART goals technique to better define your goals. Overall, I highly recommend you read the article(linked here).