CLP Data Science
Session II
How have you been?
What's on your mind?
SESSION I : CODE, STATS,
TOOLS & TECHNIQUES
SESSION II : QUESTIONS. ANSWERS. MINDSET.
SESSION II SCHEDULE
Clean up the mess as soon as the data comes in, to avoid your analysis from starting to smell.
DS2.1 : DATA HYGIENE
The world isn't normal and linear; how can we adapt our regression models to reflect that?
DS2.2 : Adv. REGRESSION
With powerful machines come powerful predictions; but how can we use them responsibly and not blindly lead ourselves into a Random Forest?
DS2.3 : ENSEMBLE METHOD
Features that come in with your dataset are often just surface representations; can we use representation learning to go beyond the surface?
DS2.4 : Adv. CLASSIFIERS
There's statistical tests for every letter of the alphabet; when aren't they just significant but essential?
DS2.5 : HYPOTHESES
What can language use tell you about a person?
DS2.6 : NLP
How do you account for irregular events like Hong Kong's hectic holiday calendar?
DS2.7 : TIMESERIES
Based on a certain sequence of meter readings, can you tell me what appliances are used in that household?
DS2.8 : DISAGGREGATION
Your code brings the numbers to the screen, now you communicate to share what the predictions mean.
DS2.9 : COMMUNICATION
We'll be looking at patterns where machine learning is half the solution, human interpretation the other.
DS2.10 : CAPSTONE
CONSTITUTION
HOW TO PREPARE?
REVIEW MATERIALS
HOW TO PREPARE?
WARM-UP CHALLENGE
HOW TO PREPARE?
GGPLOT CHEATSHEET
CAPSTONE PROJECT
PATTERN RECOGNITION
CAPSTONE PROJECT
ONE of FIVE
CAPSTONE PROJECT
PRESENT a SOLUTION
CAPSTONE PROJECT
REFLECT on FEEDBACK
CAPSTONE PROJECT
REFLECT on CONSTITUTION
How can we go faster, farther?
Made with Slides.com