Please enable JavaScript.
Coggle requires JavaScript to display documents.
Chapter 13 data science and business strategy (How does a business ensure…
Chapter 13 data science and business strategy
How does a business ensure it gets the most from a wealth of data?
Manifold
the firms management must think data analytically
the management must create a culture where data, and data scientists will thrive
Google + Amazon
Google prediction API becoming more developed and useful
predictive modeling
80s/90s/ large telephone company to reduce cost of repairing problems in the telephone network and to the design of speech recognition systems
churn prediction
capital asset allocation
Sustaining competitive advantage
can our competitors easily duplicate our processes?
Formidable Historical Advantage
Developing data assets over time; knowing buying patterns + customer behaviors
Unique Intellectual Property
novel techniques for mining data
patented
licensing
hidden methods
Unique Intangible Collateral Assets
success could also be based on intangibles like 'company culture'
Superior Data Science / Scientists
KDD cup
superior data science management
need to attract top talent
conducive environments
understanding and appreciation for data science in the workplace
understand that many problems are extremely ambiguous 'Apollo 13'
Big Red Proposal
Flaws In the Big Red Proposal
Data Unserstanding/ Data Pre: precise target variable definitions. Aligning problems of the business with the problem of data mining
Modeling: choosing the correct model, in this case a linear regression was not the correct choice in modeling a categorical target variable
Evaluation: eval should not be on the training data. Is there going to be any domain knowledge validation?
Deployment: Having a correct interpretation of the results. the idea of randomly selecting customers with regression scores over .5 is not well considered.
Make sure that it MAKES SENSE.
A firm's data science maturity
-how systematic and well founded the processes used to guide the firms' data science projects
'ad hoc' employees
hard for management to evaluate choices against alternatives
medium maturity
well defined frameworks, mimicking different aspects of the job, living off of what the mature firms are doing
Mature Organization
highest expected loss if they were to churn. Working to implement new processes, gather data necessary to judge different incentives
optimization framework