Avatar

Jas Sohi

Data Scientist

Microsoft

Jas “like Jazz🎺 music” is a Data Scientist at Microsoft based out of Vancouver, Canada. Currently, he is on the OXO Fuel team - focusing on using data to help sustain healthy enterprise customers (in terms of their usage of MS Office products).

Previously, he spent 5+ years on the modeling team at Cardinal Path - a leading marketing analytics consultancy. He was the 1st employee in the data science department and helped build it from the ground up with a focus on reproducibility.

Interests

  • R
  • SQL
  • Python
  • Statistics
  • Data science
  • Marketing Attribution
  • Customer Lifetime Value
  • Clustering/Customer Segmentation
  • Predictive modeling/machine learning

Education

  • Udacity Data Scientist Nanodegree, 2020

    Udacity

  • Udacity Data Analyst Nanodegree, 2015

    Udacity

  • John Hopkins Data Science Specialization, 2014

    Coursera

  • Bachelor of Business Admin: Finance, Management Information Systems, and Entrepreneurship, 2009

    Simon Fraser University

Skills

R

ADVANCED - 93rd percentile for R programming and 86th percentile for Data Manipulation as per Datacamp assessment

Python

INTERMEDIATE - 58th percentile for Python programming and 76th percentile for Data Manipulation as per Datacamp assessment

SQL

ADVANCED - 97th percentile for SQL as per Datacamp assessment. Use Google BigQuery on a daily basis.

Cloud Computing

Google Cloud, Azure, AWS. Google Cloud Certified (2x)

Version Control

Led and championed implementation of Bitbucket (Git) version control throughout Cardinal Path’s Data Science department

Statistics

R and Python stats packages - Experience with Descriptive Statistics, Inferential Statistics, Multiple Linear Regression, Logistic Regression, ARIMA Time Series Forecasting, Correlation Analysis, Factor Analysis, and more

Data Visualization

ggplot2 (R), Google Data Studio, Tableau

Below is an explanatory plot recently developed for a client. They asked me to forecast their estimated data storage costs in Google BigQuery:

Plot

Data Visualization

Here is another custom plot I created in R called a waffle chart. It was used to communicate the sparsity of the data and why this particular dataset was not useful for modeling (light grey squares are days with no data):

Waffle Chart

Triplebyte

Triplebyte provides technical hiring solutions. The company offers online coding tests and subsequent technical interviews to help screen candidates for prospective employers. Here are the results of a recent assessment:

triplebyte

Client feedback

“There is no marketer at [CLIENT X] who doesnt know your name, as Im always quoting you or asking for your help. You are amazing at what you do, thank you for putting up with our repetitive questions :)"

Certifications

Google Cloud Certified Professional Machine Learning Engineer

The Professional Machine Learning Engineer exam assesses your ability to: Frame ML problems, Architect ML solutions, Prepare and process data, Develop ML models, Automate & orchestrate ML pipelines, Monitor, optimize, and maintain ML solutions
See certificate

Google Cloud Certified Professional Data Engineer

The Professional Data Engineer exam assesses your ability to: Design data processing systems, Build and operationalize data processing systems, Operationalize machine learning models, and Ensure solution quality
See certificate

Adobe Certified Expert - Analytics Business Practicioner

The Adobe Certified Expert - Analytics Business Practitioner certification is the industry-recognized validation of one’s proficiency in utilizing Adobe Analytics, helping customers translate their functional requirements and measurements and proposing courses of action based on results.
See certificate

50 + Courses Completed (R, Python, and SQL)

Datacamp assessments show that my programming skills are in the top quartile of all Datacamp users.
See certificate

Clients

Apple

Built a currency exchange R script for a SQL Server database running in Microsoft Azure.

Google

On behalf of Google, presented to dozens of CMOs from around the world as a SME on Customer Lifetime Value

Past Talks

Lighting Talk at VanPy Day

Web Crawling with Beautiful Soup

A talk about using the Python package Beautiful Soup. Showcased how I scraped data from the BigQuery Web UI programmatically to save many hours of copying and pasting.

Volunteer Experience

 
 
 
 
 

Snow Angel

City of Vancouver

Dec 2019 – Dec 2019 Vancouver, BC

When there is snow and ice, getting around can be difficult, especially for seniors and people with limited mobility.

When it snows we help match-up people throughout Vancouver who have limited mobility with neighbours willing to lend a hand to remove snow and ice outside their homes.

 
 
 
 
 

Data Science Hackathon - Judge & Mentor

Vancouver Whitecaps (Soccer) Datathon

Sep 2018 – Sep 2018 Vancouver, BC

Mentored students working in R and Python on best practices as well as providing tips on presenting their results to the judges (which I was later in the day).

The aim of the hackathon was to provide an outlet for students to improve and display their skills in analytics and sport, while connecting them with professionals and researchers from sports analytics.

Bio

Jas “like Jazz🎺 music” is a Data Scientist at Microsoft based out of Vancouver, Canada. Currently, he is on the OXO Fuel team - focusing on using data to help sustain healthy enterprise customers (in terms of their usage of MS Office products).

Previously, he spent 5+ years on the modeling team at Cardinal Path - a leading marketing analytics consultancy. He was the 1st employee in the data science department and helped build it from the ground up with a focus on reproducibility.

At Cardinal Path, he led end to end projects: everything from gathering requirements, to building data pipelines, training machine learning models, and finally visualizing and presenting results. Key clients included Intel, Rite-aid, Cisco, and Universal Music, to name a few. Over the past few years, he has specialized in multi-channel attribution, market mix modeling, and predictive modeling (classification and regression).

Prior to Cardinal Path, Jas worked in sales and supply chain for a leading ecommerce startup, BuildDirect.com technologies. He honed his business acumen and worked with the COO to build up the new Inventory Department. He was one of the 1st students to complete the John Hopkins Data Science specialization where he first learned R and built a natural language model (bag of words) to predict what a user would type next. He then completed 7 projects as part of the comprehensive Udacity Data Analyst Nanodegree program taught by experts from Google, AirBnb and other industry insiders.

Nearly half a decade later, during the 2020 COVID-19 pandemic, he made the most of his time in lockdown to complete another Udacity program: the Data Scientist Nanodegree. As part of the program, he published this blog post which the medium curation team recommended to readers interested in Data Science. It was also featured on the main page of their most popular publication, the Startup.

If you haven’t noticed yet, Jas is a R ninja (this website was fully built using RStudio), a slick SQLer, and a Python pro. He’s also a bit bashful about his brilliant bash scripting skills.

On a more personal note, he’s an avid skier, hiker, and a biker. He has recently taken up tennis and beach volleyball.

Fun Facts

Canadian

I’m born and raised in Vancouver, Beautiful British Columbia, Canada

Sikhism

Sikhism, from Sikh, meaning a “disciple”, “seeker,” or “learner”, is an Indian monotheistic religion that originated in the Punjab region of the Indian subcontinent around the end of the 15th century.

Basketball

I played basketball in high school and it is my favorite sport.

5+ Years of Data Science experience

It has been over 5 years since I made one of the best decisions of my life; took a leap of faith and changed careers from Supply Chain to Data Science.

Email

I’m an advocate for Inbox zero. This is a rigorous approach to email management aimed at keeping the inbox empty – or almost empty – at all times.

SkyJump

I base jumped off Auckland Sky Tower during my trip to New Zealand in 2015.

Kaggle

I first joined Kaggle over 5 years ago. I recently led and mentored a team through a beginner competition called Learn With Other Kaggle Users

  • Classify forest types based on information about the area

Hiking

I’m a regular hiker; my favorite hike was Black Tusk near Whistler, BC

Skiing

When I’m not hiking, I’m probably skiing. My favorite local mountain in Vancouver is Cypress.

PUBG

My favorite video game at the moment is PlayerUnknown’s Battlegrounds (PUBG) an online multiplayer battle royale game on tbe Xbox.

Spicy

I ❤️ spicy foods!

Pizza

🍕 is my favorite food!