Interview.01
How do you apply traditional statistics in decision making in modern businesses?
[TIMEATTACK]DataScientist,JongbinJung_정종빈 Details
Intro
Data Scientist
Jongbin Jung
When you can perform data analysis and link it to actual business decision-making, you can be the best data scientist of the bunch.
Do you know how the Frequentist and Bayesian statistics are different exactly?
How are predictive modeling and causality analysis linked to each other?
Do you know what questions we ask first for data analysis centered on decision making
before talking about data or machine learning?
Jongbin Jung, who used to work as a data scientist and made pricing decisions at Uber Eats
tells you the answers to these questions in this class.
The top management in a company needs a data analyst
who can create insights that back up optimal decision making.
Learn about Bayesian statistics and inferences used in practical work and the secrets
including the optimal techniques that help companies make decisions
here at Coloso.
Content
Length: 32 videos
(Duration 8h 7m)
Difficulty: Intermediate
Video Details
Audio: Korean
Subtitles: English
Software Required
Google Colab (Colaboratory)
*This course focuses on statistics rather than templates and coding practice.
*If you acquire basic knowledge of statistics and Python before taking the course, you will be able to understand the content of the course more deeply.
Perks
TBD
Instructor
Jongbin Jung
Data Scientist
Hello, I'm Jongbin Jung
a data scientist.
Do you really know which statistical
technique to use
in what situation?
I'll help you intuitively understand
statistics
a data scientist must know
without using any Greek characters.
There are a very few people
who deeply worry about the problem
of empirically applying
and understanding data analysis
from a viewpoint of helping human
decision making.
Learn from that person
and become that person.
Jongbin Jung
[Current]
Data Scientist at Bolt
(San Francisco)
[Former]
Data Scientist at Uber (San Francisco)
Highlights
Class Highlights
WHY
Based on statistics that a data scientist must know, you must learn about the technique of decision making on real-world cases with high reliability in order to become a highly-paid data scientist.
WHAT
Focus on what's helpful in practical work that's centered in "decision making". I'll introduce to you in detail the confidence intervals, Bayesian, Frequentist, the significance of the model and the fundamentals of causality analysis, which many have heard of but cannot exactly define.
HOW
Show problems through intuitive examples. Learn and practice in theory or through Python code examples the statistics a data scientist must know without any Greek characters.
8 Class Exercises
Point 1. Uncertainty and Data
-
Study 1 : General Decision Making Model
Introduce a general decision making model to rationally approach the example case using as much available information as possible.
-
Study 2 : Frequentist Statistics
Among the approaches using data, solve an example problem in-depth from a frequentist point of view which is considered more "traditional".
-
Study 3: Bayesian Probability
Introduce the Bayesian approach, which seems similar to Frequentist at a glance, but the underlying philosophical position is completely different (so most importantly, the interpretation and utilization of the results are different).
-
Exercise : Which country market would you enter?
Collect the results of market research experiments from 10 market candidates and calculate which country market is the most efficient to enter. Analyze this on a statistical basis and see how the final entry strategy must be taken.
Point 2. Machine Learning Prediction and Causality Analysis
-
Study 1: Know the Basic Considerations of the Prediction Model
Know the basic considerations (train/validate/test, bias-variance) of the prediction model (and much further ML) through a primary approach using the "blindly ML" prediction model.
-
Study 2 : Exploring Basic Measures and Pitfalls
Explore some basic measures and pitfalls that approach some inevitable challenges of "causality analysis" that we encounter in decision making based on data.
-
Study 3 : An Approach Using the ML Model
Rather than introducing various ML models(random forest, NN, etc.) in a shallow manner, an introduction to a universal "approach using the ML model"
-
Exercise : Actual research analysis of example data
Predict which user is more likely to make a purchase when given a coupon, and learn about when is it more advantageous to apply coupons (for example, before and after the shopping cart), budget, target purchase rate, maximizing purchase intention, and maximizing additional revenue.
Curriculum
Curriculum
In-depth Look
Interview
with
Data Scientist
Jongbin Jung
What do you have to do to become a data scientist, how can you survive when you become one, and how can you take a leap to a step higher? To survive in this field for the next 20 or 30 years, or to become the best in this field, you need to do more than just analysis and predictions. For the last few years, I've always presented data as an insight for pricing decisions and drawn actual quantitative data every week in Uber headquarters. I will teach you the optimal technique to apply traditional statistics in decision making used in actual modern businesses.
Interview.02
I worked on deciding prices in Uber Eats, which is a delivery app used in major countries around the world.
I worked on deciding prices in Uber Eats, which is a delivery app used in major countries around the world.
Modeling, and meeting with people from different countries as making decisions in pricing is the easy part. You can discuss, "which do you believe is correct?" It's about making improved optimizations based on assumptions. Meanwhile, the parts that you cannot easily automate, in other words, the decision among the many options which you believe are correct is a difficult process. My job was to build a system where many people could decide among things that they're thinking of, judging what is the same and what is different, mechanically processing the things that are the same as groups, and things that are different to be chosen based on their own judgment. It was difficult because there were many differences between each country. In this class, I'd like to teach you the decision making process based on automation techniques using machine learning and human experience, and the basic framework for approaching complex problems where these two aspects coexist.
Interview.03
Would you like to grow into a data scientist who leads decisions?
Would you like to grow into a data scientist who leads decisions?
Right before I started this class, I left Uber in the fall of 2021 and moved to Bolt, a new powerhouse in the commerce solution field, known for its one-click checkouts. Likewise, Bolt was smaller than Uber in decision making, but there was a much larger number of products and countries offering services. There are very few people in the world who understands probability/statistics/data analysis from a viewpoint of helping human decision making instead of pure mathematics/engineering and deeply think about the problem of empirically applying them. I don't know how many people would be willing to organize a class on this. I became one of them and I'd like to teach you how. I'd like to continue to work in various private/public areas to systematically help with community decision making and policy making.
Interview.04
Data scientist or detective?
Data scientist or detective?
For a data scientist, it is helpful to put a clear distinction between things that are "theoretical" and "empirical" and address issues from the standpoint of the stakeholders with that you need to work together. When it comes to dealing with data, new machine-learning tools and coding seem relatively easy for anyone. But the difference in skills requires curiosity about the source and meaning of the data, understanding the context, and something that drives you to dig into the hidden premises. When it comes to results, the difference comes from the "detective investigation" mind which is somewhat separate from "technique" and "coding".
Required Programs
This course will use Google Colab (Colaboratory),
Please purchase and install these program(s) for an optimized lecture experience.
*The programs and/or materials will not be provided with the course.
*This course focuses on statistics rather than templates and coding practice.
*If you acquire basic knowledge of statistics and Python before taking the course, you will be able to understand the content of the course more deeply.
Recommended
Recommended Classes
Like this class?
Then check these out!
[IMPORTANT NOTICE]
*Please note that the information on this page is subject to change without prior notice.
*Would you like to learn more about our refund policy? Please check here.
*Please check our FAQ for Copyright and Proprietary Rights, User-Generated Content, Security, and Violation information.