How should you implement data science in your company?
Wanna become a data scientist within 3 months, and get a guaranteed job? Then you need to check this out !
Data science is not always an easy matter. There can be many choices that you have to make, such as:
- What problem are you trying to solve?
- Do you have data of appropriate quality?
- Whom should you hire to do this for you?
I’ve seen many mistakes happening in all parts of the process, from having the wrong data strategy, to hiring people with the wrong background, or having the wrong expectations about a problem.
I have created 2 tools with the Tesseract Academy, that can help you with this process. The first tool is the data strategy canvas. This tool is designed for those companies that do not have a data science function yet. It helps you clarify your next moves, by asking you to think of a challenge that you want to solve, and then consider all important topics around it. More specifically, you will have to think about the quality and quantity of your data, the appropriateness of the data, the cost, hiring and other matters. This tool is going to be very useful for those companies that are facing a specific challenge, which they’d like to solve through data science, but they are not sure how to start.
The second tool which is released by the Tesseract Academy is the data science project assessment questionnaire. This is a tool to use when you are working with a data scientist. Data science is often an exercise in risk management. The “science” bit means that you can never be 100% sure about how well a solution will work before you try something out. Hence, you need to be aware of the risks involved in a project, and create an appropriate plan which has to achieve two things:
- Mitigate the risk of failure (both on the data science and business side).
- Make sure that every step in the solution adds value to the company.
This can often be more complicated than it looks, since it requires a proper understanding of both the business and the problem. This questionnaire should be answer by both the stakeholder and the data scientist. A few rounds of back-and-forth should help clarify for both parties what is the best way forward.