Ahmed Elsamadisi

@christianreber @Pitch I cannot dm you either but my email is ahmed(at)narrator.ai. Please share since we are hiring backend engineers

If you're interested in hiring @Pitch talent, please DM me your email address + company name and I'll share internally.

@DavidHodge If you’re looking to answer more complex questions that ChatGPT cannot answer, we should talk. I am releasing a product soon based on narrator.ai that is perfect for that

The first ever mini-course on the Activity Schema! deepskydata.com/courses/analys…

This the Egypt-born entrepreneur was forced to lay off half of his company. Since then, he raised $13.6 million from investors like Y-Combinator. Time for a story…

🌐 Follow @moderndatastack or moderndatastack.xyz to stay updated with the latest in modern data stack. #data #moderndatastack #dataengineering #datascience #dataanalytics

🌟 The Data Modeling Paradox: How Traditional Data Models Hinder Our Ability to Ask Insightful Questions by @brittmdavis1 and @ae4ai narratordata.com/blog/the-data-…

“The modern data stack has a join problem” -Ahmed Elsamadisi @ae4ai @DataStackShow

Thanks to all the authors for producing great content this month - @ae4ai, @ergestx , @prukalpa, @imrobertyi , @bennstancil , @borisjabes & Cyril Marques (@MTL_Analytics )

For those who want to play around with the ActivitySchema I've made a repo with some code samples that use the StackOverflow database on BigQuery. github.com/ergest/activit…

It’s a holy purpose tho

There’s a universal data model that can answer any business question. It’s called language. @ae4ai gets data to tell stories in language everyone can follow. @econ4business mises.org/library/ahmed-…

I’m slowly arriving at basically the same realization many have arrived before me. For data warehouse modeling: Raw Data => ER/Relational model (3NF, Kimball, DV 2.0, Anchor, Activity Schema) => OBT view or materialized into a table) Why?

Query optimization is no longer just about speed. These days it’s increasingly about cost. Cloud warehouses will gladly run your inefficient queries and charge you for it. A slower query that spends $0.01x is often better than a really fast query that spends $1x

🤯I don't care about data, I care about making better decisions. ~Ahmed Elsamadisi @ae4ai Founder/CEO @narratorai Rarely heard #VC fundraising advice + when to trust your intuition, not data + the gift of being on purpose. 👇Tap link, grab #startup 💎💎s bit.ly/SLSAhmedElsama…

@ergestx @cfmj This is just the start. As people explore the ActivitySchema all these will be explored by @ergestx repeat activities, weird delays in time, updating data and so on. Keep following for more discoveries

@cfmj @ae4ai For multiple started, I need to figure it out. For multiple completed, you have to choose whether you attach the first one after started, or the last one after started. There’s no sessionid, you simply join in time. I’ll do a write up once I discover all of them.

I'm messing around with activity schema today. Let's see if it's really that horrible to query 😅 @ae4ai

What was the worst advice you’ve received that you now wish you never followed?

@ergestx Yeah! I think some of the details make this tricky for some people. As you continue, I would love to take some time when your ready and go over some of the nuances that arise: scale, changing your mind, multiple step funnels, identity resolution and so on.

@ae4ai I mean I can see why you'd want a tool like Narrator generating the SQL for you, since this can get pretty tedious but it's not really that bad, at least for me.

@ergestx Yeah! I am excited for you to continue. Doing a couple of activities, adding some more layers and so on.

@ae4ai I mean you have to be careful to avoid duplicates but for the seasoned analyst this should be table stakes, especially if you show them the patterns. I haven't tried chaining more than two activities together so I'll try that next. This is kind of fun actually

@ergestx Yea that looks good! I think this is great and highlights the power of consistency. And Qualify is super awesome. Wish it was on every warehouse. Do you think anyone can write these queries or do you have to be super careful?

@ae4ai This is my attempt at "first in between" for a two-step funnel analysis using a window function and the magic of QUALIFY. Seems pretty straightforward. Am I missing anything?

@JohnHMorris @ergestx Oh those are great for capturing and bringing into the warehouse. I think this is more about data prep. Most people with lots of events have a Kafka pipeline or a proxy that is using a Kafka pipeline

@ergestx @ae4ai At what point does a streaming event technology such as Flink or Kafka become relevant here?