By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
How much time should I spend?
Data science projects are easy to run for a very extended amount of time and an ever growing scope. The reason for this is very often that it’s hard to say what exactly will be the outcome of the analysis. You don’t know exactly where it will start and once you find something interesting your business counterpart will ask you the very same second: “So does that mean you can also do …..”. This causes very often that the scope keeps growing and the project never gets done. Even though the initial question is already long answered.
Agile development
In order to prevent this it is often useful to set a deadline and list the questions that you want answered. These questions you can then plan into your project in an agile way. I know it is data science but that doesn’t mean you can throw all that you know about development projects overboard. Fill your sprints with questions and tasks you want answered. This way you will stay on the track your business counterparts intended. Due to the unpredictability of data science projects it is even more important to apply good agile development. This of course goes hand in hand with the methodology of CRISP-DM, where over or within your sprints you can keep evaluating your data and model(s) against the Business understanding.
CRISP-DM
It stands for CRoss Industry Standard Process for Data Mining (CRISP-DM). The steps of CRISP-DM are as follows:
These steps are however not in sequential order. They follow the system below:
Set a deadline
Then the final deadline that you have set arrives. At this day it’s not necessary whether your project is completely done. It’s about whether the analytics that you did brought you into the right direction. Did we find our answers, is the data useful, did we find any insights. With this information you should either decide to stop the project prematurely and start putting your energy somewhere else. The other option is ofcourse, you keep on investing into this project because you have positive results and within your first timebox you shown that there is a real potential.
Do you need help with keeping your data science or analytics project on track? Or do you want us to run your analytics project for you, such that you can get a better grasp on the cost. Feel free to contact us so we can discuss what would be the best fit for you.
