Welcome back!
If you taken some of the other courses in this specialization, you might recall that
we introduced a framework called the Information-Action Value Chain.
This framework describes the sequential process that starts with events and
characteristics in the real world, and ends with an action in the market place.
In this video,
we're going to spend just a few minutes reintroducing the value chains, so
we can set the stage for the rest of the topics we'll be covering in the course.
Broadly, we can think about the practice of analytics in three parts.
Things that happen before we execute the analytical methods,
the methods themselves, and things that happen afterwards.
Lets begin by reviewing what happens before we execute analytics.
We start with something in the real world.
This could be an event, like someone making a purchase or
using a product or service, the person or product itself, or
even natural phenomenal like a thunder storm or the temperature outside.
For any of these things to be of use to me as an analyst, I need to have data.
There needs to be a mechanism that captures a physical or
digital representation of that real world phenomenon and puts it somewhere.
For every event of interest in the real world,
there needs to be a system that captures it.
We often refer to these as source systems.
For a variety of reasons, it's usually not practical for
an analyst to access data directly in source systems.
It tends to be more convenient to bring data together in some common location, and
organize it in a way that is more suitable for analysis.
Most often this location is a physical system called a data warehouse or
enterprise data warehouse, although there are a variety of methods of doing this,
both physical and virtual.
However, to actually use this data in something like a statistical
software package, a business intelligence tool or
even Excel, we need a way to extract the data we want to work with.
The extraction method will vary based on how the data is stored, but for
relational databases it usually involves the application of structured query
language or SQL.
Once we have a data set suitable for analysis, we can move
onto the second part of our value chain, actually executing analytical methods.
There are a ton of different ways to analyze data, but
broadly speaking they tend to fall into three categories.
Descriptive analytics, predictive analytics and prescriptive analytics.
If you've taken other courses in this specialization or other courses in data
analytics, chances are you've already been exposed to a number of these techniques.
But as a quick review, recall that descriptive analytics help us describe
what things look like now or what happened in the past.
The idea is to use that information to better understand the business environment
and how it works and to apply that knowledge along with business acumen to
make better decisions going forward.
Descriptive analytics can take the form of simple aggregations or cross tabulations
of data, simple statistical measures, more sophisticated applications
of descriptive statistics or advanced association or clustering algorithm.
Predictive analytics helps us take what we know about what happened in the past, and
use that information to help us predict what will happen in the future.
This almost always involves the application of advanced statistical
methods or other numeric techniques.
Finally, prescriptive analytics provides recommendations on what we should do or
what choice we should make to achieve a certain outcome.
It usually involves the integration of numerical optimization techniques,
with business rules and even financial models.
As you might imagine, an analytical project might use a combination of
descriptive, predictive and prescriptive methods to achieve a particular objective.
Once the analytical methods have been applied,
we move on to the third part of our framework.
It's really this part of the process where we will be spending most of our
time in this course.
The first step in taking our analysis to action, is to summarize and
interpret results.
Mechanically, summarizing an analysis usually involves identifying the few
charts, graphs or tables, that make it as easy as possible to see what's going on.
And supplementing those figures with a short narrative that explains
what they mean.
However, this is usually much easier said than done.
Determining how to best visualize data, can be a real challenge.
Doing this well is as much an art as it is a science.
Interpreting data analysis often requires a high degree of context and
a deep understanding about how a business works.
Once we understand what's actually going on,
we can use those insights to create a plan for how to take some action.
Often this involves setting a high level strategy, then developing specific actions
or tactics that we can take to achieve a specific outcome.
We can develop alternatives, discuss pros and cons, develop financial models, or
run a follow on analysis to help assess each option.
However, even if we're able to come up for
a plan for action that we think will have a meaningful impact,
it's likely that we're going to have to convince someone to actually do it.
We called this step of the process Delivering the Pitch.
Here, it's critical that we effectively communicate our results and
sell the merit of our proposal.
We'll spend a fair amount of time in this course talking about how to communicate
results.
But generally speaking, clarity, simplicity,
value and quality will be key ideas that guide our approach.
If we are successful, we reach the final stage of our framework, taking action.
The nature of the action depends on our strategy and objectives, and quite often
will involve another round of analysis to assess the impact we've had in the market.
So let's recap what we've covered today.
Manuel LagunaProfessor
Dan ZhangProfessor
David TorgersonInstructor
