Folding Data #26
It's been about two weeks since my last newsletter, but it feels that we live in a completely different world now. I am generally an apolitical data nerd, but I wouldn't be true to myself if I didn't share the following with you:
As Datafold, we condemn the Russian president's attack on Ukraine. Our hearts are with the people in Ukraine going through unimaginable suffering and those in Russia arrested or forced to flee for their anti-war stance, including our own colleagues and their family members in both countries.
Tools of the Week: tools to stop the deadly war
There are ways that we from the safety of our faraway homes can help. Below are three of many organizations doing good, each from a different angle:
- Nova Ukraine – humanitarian aid for Ukraine, endorsed by the Ukrainian tech community as an effective one. Nuff said.
And a couple less obvious ones:
- Meduza – a major (250K+ daily readers) bilingual independent Russian media with an explicit anti-invasion stance. Currently being blocked in Russia for telling the truth about the war sharply against the government-pushed narrative.
- OVDInfo – a Russian non-profit that helps political prisoners get out, including over 8,000 people arrested for protesting the war in just the past week.
Why those are important: due to how Russia works, the most effective way to stop the war is for more people in Russia to explicitly demand that, calling for supporting independent media and the brave voices who are being brutally suppressed by the government.
Now, take a deep breath.
We can't be indifferent to what is happening but we, the data people, also need to carry on in our jobs to make the world better with data. Keep doing good, as good will win!
So let's shift gears and talk about data stuff.
An Interesting Read: The Missing Tool For Data Exploration
BI is dead. We all know the BI space needs some bold innovation. For years, BI solutions evolved along the three primary use cases: dashboarding, self-serve analysis, and analytical deep-dives. For the past 5 years, the go-to tools for each use case respectively have been Tableau, Looker, and, perhaps, Mode. The three tools created a lot of value for teams over the years. At the same time, we all know that few things are more dangerous than trying to fit a BI tool to a use case it's not designed for. (Aside from picking Redshift as your warehouse in 2022).
We've seen attempts to merge the "what" (dashboards) and the "why" (analysis) before. Sisu approached it through multidimensional tracking of metrics across time – think data cubes + ML. This approach can provide more insight into movements of KPIs by telling us that a conversion rate drop across mobile was largely driven by iOS devices with the app version X. However, the depth of its explanations is limited by the coarseness of the data you load into it: metrics are high-level by design, and once aggregated, it's impossible to drill any further than the GROUP BY dimensions.
Motif Analytics, a new player in the BI space, sets forth a bolder vision: by working with raw events and treating them as semantically meaningful sequences (e.g. checkout happens after adding something to the cart), and turning sequences into metrics, the system is able to take the user from the big picture of a dashboard all the way to the behavior of an individual user. This not only calls for novel ways to process the data, but for a complete overhaul of the UX in BI. The team has ideas for both, and I am excited about what comes out of it!
Before You Go
Real story: I once forgot to replace my credit card in the Snowflake console when switching jobs. The company decided to take the Snowflake POC into real production...
The next bill I got was good enough for a house downpayment 😅