There are lots of sources online and in book stores that will teach a person how to do Machine Learning, Regression, Text analysis, or any other fascinating topic related to Data Science.But is that all there is?
Just know apply an algorithm, and you improve the bottom line of your company?
What if you need to justify a project? How do you demonstrate that there is a process that can be followed for data science?
Recently with some colleagues we were discussing this precise topic.
Here is the answer that we came up with:
Question:
What is the
question?
Who asked the
question?
When do they
want it?
How does it
provide business value?
Hypothesis:
Validation
Criteria
Identify the
data
Collect the
data
Transform,
Merge, Munge
Analyze and
explore
Iterate
Peer Review
Visualize and
Communicate
Generate new directions.
Produce
Finding
Finding:
TRUE
FALSE
Inconclusive
Experiment
Required
Supporting Data:
Raw Data
Chart
VisualizationI will pick a section and write more about each of these sections, but this little checklist is a beginning step in applying a Data Science process for an Enterprise.
