November Digest: It’s Okay Not To Have All The Answers
Case studies, reports, analysis, and tutorials you may have missed.
Welcome to the Data Analysis Journal, a weekly newsletter about data science and analytics.
It’s time of the month again;and I are back to advocate for analytics and entertain you with all things data across product, data science, business, and academia.
November was filled with many events, studies, and an overwhelming amount of low-quality content spread across every social media platform. We’re here to help you discover and celebrate recent case studies, analyses, and tutorials you may have missed, along with upcoming events.
🔊 Advocating for analytics
I love Mixpanel landing page:
We are often fixated on getting immediate answers at any cost. Why is churn so high? Why are people upgrading less today than in 2021? Why does retention continue to decline? Sometimes, there isn't a single answer. Sometimes, we're unable to quantify user motivations. It’s okay not to have all the answers. Keep following the user.
(This is the first—and hopefully last—publication that has taken me over 20 drafts. From venting to protesting to finally accepting and acknowledging a fascinating product analytics journey)
“The main purpose of any data analytics team is to represent trust, correctness, and confidence in data reports. Reporting revenue metrics or business KPIs demands accuracy. So, as expected, legacy BI approaches product data with an expectation of precision and correctness. However, the reality is that there is no precision or accuracy in client sessions, cookies, or events…”
🔥 November highlights
I have 5 free tickets (valued at $735 each) for the in-person NY conference to share with my readers. Reach out if you're interested in attending! (Olga)
On November 16th, Pinterest celebrated its annual Pinterest ML Day. If you missed the event, here are the recordings of 4-hour ML sessions on launching recommendation systems, multi-domain Ads ranking modeling, and more.
A few weeks ago, the Data Visualization Society (DVS) announced winners of Visualization 2023. Of 900 submissions from over 50 countries and 485 unique creators, 13 winners were chosen across 27 categories. Here are my favorites:
Build dashboards using code: Holistics, a cloud-based BI application, rolled out their v1 of dashboards as code. You can design your dashboards how you want, easily revert changes, make bulk updates, or version control with Git. Really love to see more of this (Timo).
Amplitude has appointed a new CPO. The newly appointed CPO was formerly a CPO of Tableau, having spent 13 years contributing to the growth of Tableau into what it is today. Could this mark a new beginning?
📈 Industry reports and new benchmarks
🤓 Analysis and case studies
Marty Cagan | Transformed: Moving to the Product Operating Model: The talk is focused on the product role, but in the first 10 min, he dissects pretend-only product teams who do project management but no product development, and I loved his strong take: “If you don’t measure a new feature, why did you release it anyway?” (Timo)
📖 How to write CTA copy (for paywalls) - Best practices on the CTA copy from the top 100+ apps on the App Store from, the author of newsletter.
An Analysis Of Bias Or Why A/B Testing Fails - A recap of Stanford and Airbnb's collaborative paper on A/B test setup and analysis in two-sided platforms and marketplaces.
⚙️Know your craft
When To Use Mean Or Median: The differences between types of data sources and how to figure out which data to use for analysis, modeling, or reporting.
Everyone is a CDP now: An introduction to CDPs, their types, implementation, and common challenges they come with.
What’s a data science notebook?: if you haven’t already, learn what code notebooks are, why so many data teams and developers use them, and how they work.
How to Build and Manage a Portfolio of Data Assets: I was looking for inspiration for data strategies that don’t look like they are from McK hell. Someone pointed me to this post, which has good parts - maybe a bit too admin for me. But I can pick some ideas for my approach. (Timo)
NumPy for Numpties from: an introduction to the NumPy series.
Pandas2 and Polars for Feature Engineering: A deep dive into what feature engineering is, as well as common techniques and frameworks.
The complete guide to dbt tests: this is a good “getting started” with dbt tests. It goes beyond the implementation and covers the operations of a good test setup.
❤️ Favorite publications this month
Bookmarked to re-read favorite takes
⛔ Hot Seat
Recent publications that made us raise an eyebrow.
Remember KPI tables, the most requested dashboard format:
Tableau shared a guide on how to make these in their blog: “Text tables can be turned around to build multiple KPIs without much effort.” Of yes, “effortlessly” indeed! Just a mere 10-step process involving calculations, lookup functions, cross-table aggregations, and rearranging fields. And the best part is that unless you have this guide in front of you, there is little chance you will figure this out on your own.
If you know anyone at Tableau, please pass along this recent ebook from ThoughtSpot. The self-serve analytics tool declares that “dashboards are dead”, static, outdated, expensive, and fail to deliver value. And, the dashboard development cycle should be deprecated (wait a second, didn't I say something eerily similar two years back?). Now the only hope is ThoughtSpot. Obviously.
📊 Monthly Chart Drop: 20 ways to visualize percentages
🎧 Videos and Podcasts
🍸 Drink and Mingle
Upcoming free events, meetups, talks, and webinars.
Nov 30, NZ: dbt: Modern Data Stack Chat Wellington
Dec 1, NY: DSS: Applying AI&ML To Finance&Tech
Dec 1-3, virtual: PyLadiesCon 2023
Dec 5, SF: Data Meetup with Snowplow
Dec 7, Palo Alto: Product led Sales Meetup
Dec 12-15, Prague: The 13th Annual PostgreSQL Conference Europe
Dec 13, virtual: Google Cloud Applied AI Summit, AMER
Thanks for reading, everyone!
Olga + Timo