Data Science is not Rocket Science!

                                      

Topics To Be Covered: 

• Introduction to Data Science


• History & Evolution 


• A typical day in a Data Scientist’s life.


• Application of Data Science

Data Science is all about using the right mix of data, techniques, tools, and people to create a business value for an organization.

A lot of people have misconceptions about Data Science that it is about learning tools and technologies, but you also need Mathematical and statistical knowledge. You learn this for application perspective more than formulas. Third skills is business knowledge, their values, business objective, and the data they have.

                                                                                

                                                          Data Science is at the nexus of domain, Math & tech skillset. 

ML, AI, Statistical Modelling, Business Analytics are all subsets of a larger super set of Data Science. 

Let’s discuss it with an example, as we know Covid situation is considered seriously so the purchase of masks is large, but when covid cases will decrease then there will be an exponential decrease in the purchase of masks.

Covid Cases       Mask Purchased

50,000                  5,00,000

50                         500 and so on

So, using this data many businessmen are producing masks, and also altering the workforce. This is a real-life situation example for Data Analysis.

                                    

Your learning will never be complete if you only know technology, or even mathematics, So, you need to learn all three skills, Technology, Mathematics, Business Knowledge.

4M of Data Scientist

• Mindset: We need the right mindset for approaching business problems    




• Data Thinking: Whenever a dataset is given to you then with that which problem can you solve. This is another way around. 


• Scientific Reasoning: No conclusion can be drawn based on data, or evidence. 




• Mechanics: We should have enough knowledge/ experience of Market/Business




• Business/ Domain: Knowledge of solving business problem which model


• Market Dynamics: knowledge of changes in market with respect of your model.




• Model: We should be aware of various models and their application




• Application: Learn various model, application


• Concept: Learn concept, which model solves which problem 




• Method: We should be updated with new tools and technology




• Tool: Use tools, learn tools, R, Python, Tableau


• Dataset: various forms of dataset, structured and unstructured

If you have the right mindset then you can take out insights from data, and this will increase the pay scale as we go upward.

History & Evolution

Before the 19th century: There were no proper rules or regulations, so that managers could track how workers were performing. If workers are not efficient then business efficiency was minimum.

Business Efficiency ….                  Data Generated  .

Introduction to Scientific Management

Frederick W. Taylor initiated time-management practices to improve business efficiency.

He used different types of shovels to check the amount of sand displaced in given time.


                                                       

He recorded all the observations, so that it can be used by his organization and others as well.

Frederick W. Taylor mentions, “In the past, the man has been first, in the future the system must be first… The first object of any good system must be that of developing the first-class man.”, so it means, if man is efficient we are dependent on people, but when the process is laid for an organization then we are less dependent on people and more on the procedure.

For eg, Wherever you go to MCDonalds then the process is same for making burgers, and so the chain of MCdonald's are not dependent on the people, rather on the process.

This was “Standardizing the process”.

MCDonalds, standardized the process by dividing the complete procedure of making burgers into different people rather than one person.

Business Efficiency …….                                Data Generated  ...

The Dawn of Business Analyst 

Then, comes the time, when Business Analysis started to rise, which defines the procedure for approaching any business problem effectively. 

Business Analysts understand the risk bottleneck, and document it, so that it can be reviewed later on.

For eg, based on a credit card company, there will be marketing team,  

                                  

Steps are as followed:

• Creating a business architecture


• Preparing a business case


• Conduct risk assessment


• Eliciting requirement


• Analyze a business process


• Identify efficiencies/ bottle-neck


• Documenting requirement

Business Efficiency ……...                    Data Generated  .....

Business Intelligence

Business Intelligence came into acknowledgment, which helped in presenting data or analysis reports using visuals to Business Users/ Stakeholders.

Business Efficiency ….....                            Data Generated  ……..

It is time for Data Scientists Phase1 when Data Scientists are experimenting, and they have more skill sets as compared to the past Business Analysts. Skill set has more as compared with business Analyst.

Business Efficiency ……..                             Data Generated  …………….…..

Data Scientists Phase2: It is believed that in the future, the age of Data Science will come where there will be no humans in the loop, complete automation.

Business  Efficiency ……………..…..           Data Generated  …………………..

A Typical Day in the Data Scientist’s Life

• Problem Definition: The main problem is addressed
  
  
  
  • Problem Identification: If there is any challenge any team is facing, understand the problem.
  
  
  • Problem Categorization: Some problems can be solved by providing reporting or dashboarding, eg, how much lead is generated by marketing team, in exploratory data analysis, you can suggest the improving the condition, Prediction is done for providing the trends in future, or in Prescription, with prediction the procedure is also discussed.
  
  
  • Problem Prioritization: Prioritization is done accordingly to revenue generation
  
  
  
  


• Data Management- Collecting, verifying & governing: Collection of right, accurate data
  
  
  
  • Data Sourcing: Understanding data and usage
  
  
  • Data Quality: Data should be complete, coherent, consistent, and correct, eg; Users either do not provide complete data, or they just put wrong data. Data should be reliable.
  
  
  • Governance: Right people get the right data at the right time in the right format to make the right decision.
  
  
  
  


• Reporting/ Dashboarding: Presentation of data using visuals
  
  
  
  • Reporting: Reports are generated timely. Ad Hoc: one time report, Standard: Regularly generated.
  
  
  • Dashboards: Data which is represented live.
  
  
  
  


• Insights Generation: Using mathematical models for doing the actual processing
  
  
  
  • Model Selection: Select model, which model solves your problem
  
  
  • Model Training & validation: Various models are created and checked which one is best for your use case.
  
  
  • Model Implementation: Model is implemented and a solution for the problem is provided. Deploy in a production environment.
  
  
  
  


• Insights Distribution: Sharing reports and conclusion
  
  
  
  • Analytical Storytelling: Telling stories with data and providing the solution for the problem
  
  
  • Change Management: Shift to data driven decision making
  
  
  
  


• Benefits Realization: Sharing the outcomes from the analysis
  
  
  
  • Attribution methods & reporting: Agreement on attribution method, periodic reporting to management.
  
  
  • Leading Indicators: Conversion, acquisition, periodic, satisfaction rates.
  
  
  • Lagging Indicators: Revenue, cost, and profit
  
  
  
  

Applications Of Data Science

• Telecom
  
  
  
  • Predict and prevent users to move to competitors companies
  
  
  • Customer-Centric pricing plans
  
  
  
  


• BFS
  
  
  
  • Risk Management
  
  
  • Up-sell and cross-sell
  
  
  
  


• Sports
  
  
  
  • Predict the game outcome
  
  
  • Improve player’s performance
  
  
  
  


• Life Sciences
  
  
  
  • Predict the development successes
  
  
  • Identify early adverse reactions
  
  
  
  


• Insurance
  
  
  
  • Claim Analytics
  
  
  • Reduce fraud and abuse
  
  
  
  


• Retail
  
  
  
  • Promotion Planning
  
  
  • Redefine shopping analysis
  
  
  
  


• Energy Utilities
  
  
  
  • Monitor energy usage pattern
  
  
  
  


• Entertainment
  
  
  
  • Digital Customer analytics
  
  
  • Recommendation
  
  
  
  


• Healthcare
  
  
  
  • Evidence-based medicine
  
  
  • Identifying patterns and deviations revealing clinical insights
  
  
  
  


• Automotive
  
  
  
  • Self-driven cars
  
  
  • Fraud with vehicles during service