298 | Big Data Simplied
The schema corresponding to the data structure in Figure 12.3 is as follows.
custId: Customer identity number
custName: Customer name
custAddress: Address of the customer
custLocation: Customer location
custPhone: Customer cell phone number
productBrand: Brand of the product
productCategory: Category of the product
count: Count of the product brought
dop: Date of purchase
Then specic analytics have to be applied on this data pool to come up with customer lists and
a discount slab to which they can be mapped to. In the next section, we shall dissect the prob-
lem and come up with a solution approach which can be implemented using different tools of
BigData.
Did You Know?
Customer churn means unhappy customers stopping to do business with a company and
moves to a competitor company. Customer churn is a major headache for all companies. Ina
survey conducted by CMO council, the global marketers told that customer churn signi-
cantly impacts the following parameters.
• Business performance through revenue loss.
• Reduced profit.
• Greater marketing and customer re-acquisition costs.
Yet nearly 67% of surveyed respondents said that they have no system for identifying poten-
tial customers who may churn and retaining them!!!
12.3 NRT ANALYTICS: SOLUTION APPROACH
The problem laid out in NRT analytics can be broken down into three main pieces as listed below.
a. Accept the streaming data input.
b. Persist the data in some data store.
c. Apply analytics on the persisted data.
To simulate the streaming data input, Kafka Producer can be implemented in streaming batch
mode using Java. It means the Kafka Producer will be enabled to produce newer record at regular
intervals of time, say after every 2 minutes. For accepting streaming data input, we can plan to
M12 Big Data Simplified XXXX 01.indd 298 5/13/2019 10:02:09 PM
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.