Thanks for being here!
Announcement(s):
Now you can access 90 West’s consumer transaction data via our Insights Platform (powered by Exabel).
Theme that emerged in this week’s email is … the market is developing a greater appreciation for the value of data. This starts with companies getting the arms around the value of their internal data & the benefit they can see from simply organizing what they’ve already got.
Quotes:
“…all that progress in algorithms means it's actually time to spend more time on the data,” - Andrew Ng (discussing progress in Artificial intelligence)
News Articles
Podcasts
Cool Charts
Final Thoughts (The Prince)
#1 – Sara Brown of MIT Sloan published Why it’s time for 'data-centric artificial intelligence'. A conversation with Andrew Ng. July 2022.
My Take: I try to follow whatever Andrew Ng is saying … surprised I didn’t catch this July article sooner. AI progress is severely hampered by bad data. “Data-centric AI” is the discipline of systematically engineering the data needed to build a successful AI system. Basically, solving the garbage-in = garbage-out problem. These are huge problems to solve. One suggestion in the article is to tackle smaller projects or single vertical, create standards. I compare it to a CRM (Customer Relationship Management system)…there are good CRM systems that do OK for all firms, but CRMs designed for the nuances of specific verticals are better. At this time in the maturity of the AI/data market, it is tough to target anything but the broadest markets, or multipurpose AI systems as described in the article.
#2 – Chelsea Wilkinson of Data Diligence published Data is a risk to mitigate & opportunity to realize in all M&A. October 2022.
My Take: The awareness of data value as an asset is increasing. Data represents value that is largely left off any balance sheet accounting. I would argue any company, especially a company that may be acquired, should complete an “audit” of data assets and understand the value. Going forward, the company’s proprietary data will be the most valuable asset.
#3 – Shivina Kumar of Spectre Data published Working with alternative data? Here's how to prepare. August 2022
My Take: THE hardest part of selling alternative data is getting the data in order. It is not glamorous. It takes time. Tools designed to help are gaining traction and getting better. Most of the “cost” associated with dealing in data are “hidden costs” that steal momentum before you even get going.
BONUS: Chad Sanderson & Adrian Kreuziger published a series on Data Contracts. An Engineer's Guide to Data Contracts - Pt. 1. October 2022. Almost too much in this article for me to summarize effectively. Very good read.
BONUS 2: Thaila Barrera of AirByte blog Reverse ETL Explained. August 2022. Good place to start for non-technical people who want to understand concepts like ETL, ELT, Reverse ETL.
BONUS 3: Alex Izydorczyk published Simple, Fast, and Transparent Data Sales. October 2022. Principles for data monetization.
#1 – Seattle Data Guy has a YouTube Channel with 43k subscribers. Good mix of short 5-15 minute instructional videos & longer deep dives on various products and topics.
My Take: I checked out and will share the video Vocabulary for Data Engineers - Data Engineering 101. Good overview of a common data terms.
Highlights (15-minute run time):
Minute 00:30 – DAG (a set of tasks that need to happen in a particular order)
Minute 01:30 – Data Pipelines, ETL, ELT, IPaaS
Minute 04:45 – Data Warehouse (rigid schema) & Data Lakes (less rigid)
Minute 07:45 – Tables (database vs data warehouse; facts & dimensions)
Minute 09:30 – Slowly Changing Dimensions (SCD – a way to track historical data)
Minute 12:30 – SLA (Service Level Agreement)
Source: Raconteur’s Embracing AI.
The below image can be tough to read…so click here for the image in a better format.
Key takeaway for me = China & India have big leads in AI deployment.
BONUS: From AirByte blog about ETL and reverse ETL. Data Hierarchy of Needs
BONUS 2: Reminds me of the MIT / Andrew Ng article cited above.
Source: I came across an interesting podcast series (not related to the world of data). I’d recommend you check out The Prince. Published by the Economist. This is an 8-part podcast recounting the rise of China’s leader Xi Jinping. Fascinating.
Most notable ideas:
The CCP (Chinese Communist Party) wants to avoid civil unrest. If you make it possible for a person to provide food/shelter for their family & have the hope that their children will have a better life than them, that is the most direct path to a content society.
The CCP is focused on avoiding civil unrest by providing the best life for the most people. One downside of this is they have demonstrated a willingness to do this at the expense of the minority.
I am fascinated by the Mao “Great Leap Forward” period in China (books recommendations welcome). Xi Jinping grew up during that period and it had a profound effect on him and his thinking, but not how I would have thought.