Data Science projects consist on different steps that sometimes require different skills. Here is how I can help you in each of those steps or in the project as a whole.
Everyone wants to have a data driven approach, but many times excecutives don't really know what that means. I help to translate the business needs into data analysis projects. How? reviewing already available reports, interviewing the right people, asking relevant questions, analyzing what competitors do or state of the art for the industry. During this stage I also determine if the results of a model could be applicable to business operations.
Many organisations gather a lot fo data, but sometimes it's not the one that's needed or the quality is not the best. Based on the business needs, I define the data that is needed, map it to availability and analyze the quality and completeness. Depending on the maturity of the organization, it might also be needed advice on how to collect the data and structure it. I work on the conceptual design. After that Data Engineers can take over.
Once the data is made available, I conduct a first explorative analysis to determine the richness of if. Form hypothesis and visualize the information. Present the results to validate with business owners.
This is known as Feature Engineering. Many times Data Science projects fail because the right variables where not created. Together with the Data Scientists, we work on the models to be used and what type of variables help.
If after the above steps we determine that advance analytics could improve business performance, we get to the real data science. In this stage I let the Data Scientists to do their own work. I contribute by managing the project and checking with developers the status. You don't have data scientists in your team? I can find one for you.
After the model has been completed by data scientists, I work with them evaluating the results, defining the value and the business aplication. Many times the results of a complicated model cannot be implemented in real life, so that's why a this point the domain knowledge plays a huge role.