Machine Learning: 2 Books in 1: Machine Learning for Beginners, Machine Learning Mathematics. An Introduction Guide to Understand Data Science Through the Business Application

Stage I – Business understanding
The goal of this stage is to gather and drill down on the essential variables
that will be used as targets for the model and the metrics associated with
these variables will ultimately determine the overall success of the project.
Another significant objective of this stage is the identification of required
data sources that the company already has or may need to procure. At this
stage, the two primary tasks that are required to be accomplished are:
"defining objects and identifying data sources".
Deliverables to be created in this stage
Charter document – It is a "living document" that needs to be
updated throughout the course of the project, in light of new
project discoveries and changing business requirements. A
standard template is supplied with the TDSP "project structure

---

definition". It is important to build upon this document by adding
more details throughout the course of the project while keeping
the stakeholders promptly updated on all changes made.
Data sources – Within the TDSP “project data report folder”, the
data sources can be found within the “Raw Data Sources” section
of the “Data Definitions Report”. The “Raw Data Sources”
section also specifies the initial and final locations of the raw data
and provide additional details like the “coding scripts” to move
up the data to any desired environment.
Data dictionaries – The descriptions of the characteristics and
features of the data such as the "data schematics" and available
"entity-relationship diagrams", provided by the stakeholders are
documented within the Data dictionaries.
Stage II – Data acquisition and understanding
The goal of this stage is the production of high quality processed data set
with defined relationships to the model targets and location of the data set
in the required analytics environment. At this stage "solution architecture"
of the data pipeline must also be developed which will allow regular
updates to and scoring of the data. The three primary tasks that must be
completed during this stage are: "Data ingestion, Data exploration and Data
pipeline set up".
Data ingestion
The process required to transfer the data from the source location to the
target location should be set up in this phase. The target locations are

---

determined by the environments that will allow you to perform analytical
activities like training and predictions.