1. Companies collect a lot of data, but how much do they actually use?

55% of the data collected is “dark data”—data that’s never used.

2. The future of data engineering

Chris Riccomini discusses what to expect in the next few years in data engineering and four areas he anticipates will face major change.

+ Check out the data engineering sessions at the O’Reilly Strata Data Conference in New York. And the data engineering resource center on O’Reilly learning.

3. These companies claim to provide fair-trade data work. Do they?

There’s a lot of human labor in building AI systems, mostly tedious cleaning, categorizing, and labeling of data. Many of these invisible human workers live in India, Kenya, Malaysia, and the Philippines. Companies like CloudFactory, iMerit, and Samasource promise datasets provided by workers who are well paid and cared for. But standards vary widely.

4. What is decision intelligence?

“Decision intelligence is a new academic discipline concerned with all aspects of selecting between options.” Cassie Kozyrkov explains.

+ Cassie Kozyrkov will give a keynote at the Strata Data Conference in New York, September 23–26.

5. Introducing the funneljoin package

“Have you ever had a ‘first this then that’ question? For example, maybe you’re an e-commerce business and you want all the times people clicked on an item and then added it to their cart within 2 days, or the last page they visited before registering.” If so, you may want to read up on the funneljoin package for R.

Small-group, hands-on training at Strata in New York

Strata’s two-day intensive, expert-led training courses give you hands-on experience in the technologies you need to succeed—they cover everything from TensorFlow to recommendation systems to serverless data apps. And the class size is kept small so you get individual attention. But that means they fill up fast. So reserve your seat at the courses that interest you ASAP.

6. Modeling conversion rates and saving millions of dollars

This post clearly shows how to analyze conversion rates and conversion rate changes over time and estimate lifetime value—without substantial delay. It includes a description of Convoys, a Python package that you can use to fit the models described. Recommended.

7. Python is eating the world

“How one developer’s side project became the hottest programming language on the planet.”

8. I studied machine learning every day for 9 months, then got a job.

Daniel Bourke taught himself machine learning five days a week and drove Uber on weekends to earn himself an ML job. Here’s how he did it.

+ Looking for a place to start your own self-designed master’s program? Check out O’Reilly learning.

What do you get with the O’Reilly AI Conference Expo Plus Pass?

You get a lot for just $145. If you’re not ready to commit to the full O’Reilly AI Conference in San Jose experience, take a look at the Expo Plus Pass. You’ll be able to test-drive new tools, compare products, attend sponsored sessions, meet with speakers and authors, get access to 90 days of O’Reilly online learning (a $117 value), and much more. It’s a great opportunity to see what’s new in AI tools and technologies, network, and get a taste of the O’Reilly AI Conference.

9. Progression of a data scientist

“The career trajectory of a data scientist depends chiefly on how much impact they are able to have.”

10. What every data analyst and data scientist should be able to do