Action
The next step in the data lifecycle is action.
Knowledge for the sake of knowledge is a noble
pursuit and valuable in it’s own right.
However, in order for data to be valuable
to our business, it must lead to some form
of action.
So, in data science we use the results of
our data analysis to take an action of some kind.
This stage begins with making a decision on
what action to take based on the data and
our analysis.
This may also include making the decision
not to act (i.e. choosing inaction based on
our analysis).
Next, we take the appropriate action in order
to affect positive change.
For example, we might approve a customer for
a loan, recommend a product on our website,
or change a business process.
Finally, this stage ends with an outcome which
is either positive, negative, or resulted
in no change at all.
We need to observe and record the outcome
of our actions as data, because we’re going
to use this outcome in the final step in our
process.
Action from data can take on many forms.
For example:
We can make a decision based upon our own
data analysis and then act upon our decision.
We can communicate our findings to a wider
audience in order to encourage others to take
a specific action.
Or we can automate a decision-making process
with a computer so that the action happens
automatically when the right pattern of data
are observed by the machine.
No matter who is performing the action, we
always want to make sure that we’ve chosen
the best action given the data.