Data Science and Aritificial Intelligence
Notes from Live Session
Data is the new oil
Impact
Impact on Banking
- Querying the data is huge problem for Banking. Here, ML, AI, Data Analytics is used. - Banking
Impact on Retail - Snapdeal
- Data Science has moved to central role from support role
Impact on IT Industry
How things have changed over time in Analytics
- what happened (descriptive)
- how it happened (inferential)
- what will happen (predictive)
- how we can make it happen (prescriptive)
- Cloud Modernisation - https://www.snowflake.com/
Impact on Retail - BigBasket
-
Retail customers want
- great range
- convenience
- price
- Not all customers will focus on all three things at all times. It depends on the situation and timing.
- Personalization is not a new concept, just that number of customers is huge now.
- The repetitive process should be as simple and quick as possible so that customer can spend time on discovery.
Data Element - What exactly is it used for
Healthcare
- medical record is dense
- try to make patients life better
- advance research
- finding doctor and information
- onboarding data points
- understanding data we already have
Finance
varies from client to client
-
retail
- spend behavior for retail
- savings not that important
- spend should happen from your app.
-
broking house
- trade data, portfolio
- banks earn max amount from this category
-
other aspects
- consumer behavior
- risk fraud detection (KML)
- banking is complete tech driven
- AR (Augment Reality) is being used in Banking a lot
Retail - Snapdeal
difficult to point out one specific data point
- customer experience
- repeating customer
- how much time does user spends - recommendation model (#1)
- negotiate with sellers and customers
- price optimization - pricing model (#2)
- delivery - predictive model
- image processing - de-duplications
Information Technology and Services
- client doesn't come up with tech spec
- most important is domain experience, then followed by technology
- harmonize the data so that data can be used by anyone
- self-service
- UI/UX
- what am I able to understand through the dashboard
- have design thinking sessions, followed by user personas
- provide insights that are accurate, trust in data
- adaptability
Grocery - BigBasket
- to customers it doesn't matter what goes in pot, what matters is what comes out
- retention
- ease of use
- what are products they are likely to run out of
- smart basket makes shopping experience better
- frictionless process
Is Upskill necessary
- In short, yes
- Learning has become easier
- Learning what works well in what situation
- Follow through hierarchy
- become master of few technologies
- writing code is important to improve brain
-
determination
- how much you want to learn
- don't compare with peers
- What you want to achieve in life
- What have I learned today?
- don't stop learning
- have to be realistic when switching jobs
- open minded
- keeping mind sharp
- learn core skill set and enough different skill sets to know what will work
- how you can add value in the team - data science is not one person job, it's a team effort
Models I should study
- XGBoost
- RandomForest