Ultralearning Product Data Science

Author(s): Harsh Darji

Originally published on Towards AI.

Data Science

The minimal effort maximum outcome way

Happiness is not pleasure. Happiness is the expansion of possibilities — Scott Young

I often wonder what is one thing I am most passionate about, I am aware I like fitness, sports, e-commerce, and advertising space but I cannot come up with one answer. Is it because I know I am not good at it or am I scared to actually make a career out of it? I thought completing grad school and finding a job will help me to find my passion, it has definitely made me happy but I am far from finding my passion. I tried digging into the work of some of the people I admire to find an answer and I came across Cal Newport’s book “So good that they cannot ignore you” and my mind was literally blown. I was trying to find my passion and within the first 10 pages, Cal said Passion is overrated and the key to be good at anything is to force yourself through the work, force the skills to come and that’s the hardest phase.

I feel like your problem is that you’re trying to judge all things in the abstract before you do them — Cal Newport

After understanding the passion hypothesis and pairing it with the learnings from Scott Young’s book “Ultralearning” where he emphasizes we can learn anything irrespective of our background, I have decided to work on building my career capital which is to develop skills in product data science. This is the Ultralearning Product Data Science program that I have prepared for myself to follow for the next 100 days (01/03/21–08/06/21) Check out my notes of Ultralearning and So good that they cannot ignore you to comprehend.

Why Product Data Science?

I have a Master’s degree in Applied Data Science and I am quite good with algorithms, coding, and statistics and instead of building my career capital in an entirely new industry (Finance/Marketing/Arts), I have decided to leverage my strength and upskill! I know data scientists get paid well that is another important factor I took into consideration and I would be lying if I say money is not important to me. Here are other reasons why I want to develop my career capital in product data science:

• 70% of jobs in Data Science are related to Product Data Science
• Less coding more analytical work. I am good with stats and decision-making!
• I am curious and I tend to ask a lot of questions so I believe I can think of questions that the rest of the company has not thought of
• Love to research about international markets, growth strategy, consumer behavior
• Easy transition to Product Management

What skills I will be developing?

I read everything that I could find about product data science over the internet and also reached out to a couple of data scientists in my network to learn more about product data science, skills needed, projects I should be working on, interview structure, and the resources I should be using. The product data scientists work closely with product leaders to run experiments and test out different product features. Some of the problems that they work on are:

• Logging and Instrumentation — Identifying Data Gaps. Establish good relationships with engineers
• Push Notification Analysis — How many users are eligible for push notifications? across user segment? across clients? What are the tap rates of different push notification types?
• SMS Delivery Rates — How do we calculate Twitter’s SMS delivery rates across different carriers? Are our delivery rates in emerging countries poorer? How can we make them better?
• Multiple Accounts — Why do certain countries have a higher ratio of multiple accounts? What drive people to create multiple accounts?

There are plenty of resources available online and I have carefully selected some of them based on my style of learning. Following are the skills I will be developing and resources I will be using for the next 100 days.

Product Sense

• Stellar Peers
• Cracking the PM Interview by Gayle Laakmann McDowell and Jackie Bavaro
• Decode and Conquer by Lewis C. Lin

SQL

• Leetcode database problems
• HackRank SQL problems
• Interview Query SQL problems

Statistics

• Online Stat Book
• Practical Stats for DS

Experimentation

• Udacity course
• EXP Platform
• KDD Papers
• Ron Kohavi’s book — Trustworthy Online Controlled Experiments
• Company Tech blogs — Airbnb, Uber, Facebook.

I am fairly good with coding in python and basic machine learning so I am not scheduling a separate time to learn algorithms and solving a take-home challenge will help to brush up on these skills. If you want to check out all the resources I collected to learn more about product data science, check out my notes

How I will be building these skills?

Instead of developing each skill individually or working on a full-stack project, I have decided to do Interview based learning. I am using the directness approach which Scott mentions in his book. The principle of directness asserts that it’s actually while doing the thing you want to get good at when much of learning takes place. So, I want to get good at interviews and I believe directly solving the interview questions will help me to do so. There are plenty of interview questions available over the internet, but here are some of the resources I will be using — InterviewQuery and Glassdoor.

I have also found opportunities within my company to develop my analytical skills (SQL, data viz, business case, scripting). My company is an e-commerce company so there are a lot of opportunities to build product skills. My goal is to be a part of at least one experiment so that I can understand the nuances of A/B testing and experimental design. When we learn new things, we should always strive to tie them directly to the contexts we want to use them in.

Writing brings clarity to me and I believe sharing my solution to product case questions will help me to receive feedback and I can improve. I will be writing regular blog posts about product-related questions asked in the interview like:

• Let’s say that you work for a SAAS company. A PM comes up to you and tells you that users that come from Country X convert from a free trial to paying at a much higher rate than users from Country Y. What would you investigate further to determine why this was happening?

I am working a full-time job so I cannot invest a lot of time on this project but I have scheduled for 20 hours a week. I will be spending 2 hours a day on every weekday (1 hour for 2 SQL problems and another hour to work on 1 product case question) and 10 hours on the weekend ( Read reference books, solve take-home challenges, read papers and summarize)

I am very excited to try this ultralearning program and see what I can achieve in the next 100 days. I will be sharing a weekly plan and the learning progress on my blog. Reach out to me if you want to follow along! If you are interested in learning what I am up to in 2021, please subscribe to my newsletter or follow me on Twitter

This blog was first published on harshdarji.com

Reference

1. https://towardsdatascience.com/how-to-land-a-data-scientist-job-at-your-dream-company-my-journey-to-airbnb-f6a1e99892e8
2. https://towardsdatascience.com/how-i-got-4-data-science-offers-and-doubled-my-income-2-months-after-being-laid-off-b3b6d2de6938
3. https://medium.com/@rchang/my-two-year-journey-as-a-data-scientist-at-twitter-f0c13298aee6

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI