Schedule

10:00 AM
Grace Tang
Senior Data Scientist
Netflix

Lessons Learned in My First Decade as a Data Scientist 

Data science is a relatively young and rapidly-evolving field, and unlike many other professions, there is no well-trodden path to success. In this talk, I will share some things I wish I’d known when I started in data science almost ten years ago, highlighting challenges and lessons I learned along the way, with a focus on soft skills.

10:30 AM
Tim Eck
Insights Lead
Google

Modeling Media Effectiveness with Lightweight (Bayesian) MMM

Interested in understanding the impact that each media channel has on your KPIs and how this can help with budget optimisation? Join us to learn how you can use Marketing Mix Modelling with Google's Bayesian open source library, LightweightMMM. Taking a Bayesian approach to MMM allows an advertiser to integrate prior information into modelling and construct hierarchical models, using breakout dimensions such as geography.
This talk will walk you through the library that allows users to easily train MMMs and obtain channel attribution information, and includes capabilities for optimising media allocation as well as plotting common graphs in the field.

11:00 AM

Tanya Tandon

Data Scientist
Google

Storytelling via Data

Storytelling is everything. It's how we learn. It's how we connect. It's how we ship products. This talk will cover some fundamentals on storytelling via data and why it is critical for your success in Data Science. 

11:30 AM

Amit Bilgi

Data Scientist
Stripe

Out of sample testing for time series data 

Time-series data (or any data with an explicit order) violates a core assumption of many machine learning algorithms and techniques- that the data are independent and identically distributed or IID. This may result in a variety of challenges when determining if a model is good, as traditional out-of-sample testing may result in leaking information to the test set. In this talk, we'll go over a variety of real-world examples of modeling time-series data and some best practices for model evaluation.

12:00 PM
Lunch Break

12:30 PM

Mahtab Dhaliwal

Data Scientist
Shopify

Biases in Data Science and best practices on how to handle them

Bias can be introduced in any stage of data processing - from collection to visualization. As data scientists, we turn data into information. Being able to identify sources of bias and knowing how to account for that adequately, brings the resulting insights closer to the source of truth. In this talk, I’ll be covering bias in sampling, reporting, algorithms, and interpretation - each with examples, cautionary tales, and statistical methods to mitigate influence of each bias type.

1:00 PM

Mengying Li

Data Science Leadership
Notion

Amplify Your Impact as a Data Scientist

A great data scientist is not necessarily the most technical. To be successful, you must go beyond your technical knowledge and learn to proactively identify business problems, communicate effectively, collaborate with others, and focus on outcomes that matter and drive the progress of a project on your own.