Catalog & Cocktails: The Honest, No-BS Data Podcast

Takeaways with Theresa Kushner

Where are the semantics in the data dictionary? w/ Dan Bennett

Machines and people. Why can't we just speak the same language? The truth is we can, and doing so could make life demonstrably better for data scientists. Yet here we are, living in a world of rows and columns that few people outside of the data owner understand.Join this weeks episode of Catalog & Cocktails as hosts, Juan Sequeda and Tim Gasper with special guest, Dan Bennett, tackle semantics and how to get everyone -- machines and people -- on the same page.Key Takeaways:[00:01 - 02:47] Intro & Cheers[02:49 - 04:57] If you were the picture for a word in the dictionary, which word would it be?[04:58 - 08:35] The Greatest Sin of Tabular Data[08:40 - 11:02] Examples of semantics missing inside of tabular data and their utility[11:03 - 12:39] Adding context and profiling data[12:43 - 14:55] How are constraints and semantics being defined, and what is a scaleable approach?[15:02 - 16:24] Data producers and enriching data[16:30 - 17:58] Enrichment that travels with the data[18:00 - 20:14] What are the tools we use, the data dictionary, and standardizing[20:16 - 22:34] Metadata and the bridge to the semantic world[22:36 - 24:57] Innovation and Dan's thoughts on relational model table relationships[24:57 - 28:09] Solving the same problems over and over again[28:12 - 30:19] Network effect, the marketplace of ideas and social spheres[30:21 - 33:56] Diving into the network effect and the semantic world[33:58 - 36:48] Why redefine if an option exists that can be used, and thoughts on simple ideas being the best solutions[36:52 - 40:11] Figuring out supply and demand curves for S&P Global[40:12 - 44:34] The business value of data and data literacy in accurate findings[44:44 - 46:19] Advice to data leaders and vendors[46:37 - 50:14] Lightning Round[50:29 - 56:48] Takeaways[56:52 - 59:44] Three final questions

Takeaways with Dan Bennett

Put the Business in charge of their own data w/ Gabi Steele and Leah Weiss of Preql

Data and business teams become a convoluted intersection, and when they struggle to communicate, it leads to bigger problems than awkward water-cooler talk.So what comes first? Translation? Data literacy? Company culture? The chicken? The egg?Co-founders of Preql, Gabi Steele and Leah Weiss, join hosts Tim and Juan to discuss how to put the business in charge of their own data and how this leads to the answers AND massive alignment between data and biz teams. Key Takeaways [00:05 - 04:15] Introduction and Cheers to Tim[04:17 - 06:50] What is the wildest or weirdest thing you've seen while stopped at a red light?[06:55 - 10:29] Putting businesses in charge of their own data[10:29 - 12:04] Inviting the business into the process and building community[12:05 - 15:21] Developing a curriculum for data, teaching SQL and data visualization[15:22 - 19:17] Refining the model and curriculum, tailoring the fit[19:17 - 21:41] Identifying the people who wanted to be part of data modeling[21:43 - 22:55] Teaching interested parties another way to handle data, using an application process to find people[22:57 - 24:17] Stewardship in handling data[24:17 - 27:04] The process of engaging and completing data modeling: a brain for architecture, python, and sql[27:05 - 28:32] Interesting problems to solve, and you have to be creative to get there[28:36 - 30:40] Understanding technical debt and how to support engineering teams, skills for good data modeling stewardship[30:42 - 33:18] DBT and building robust analytics teams[33:30 - 37:35] Tooling solutions for helping business be in charge of their own data[37:42 - 38:52] Focus on the technical side, and a no-code semantic layer[38:53 - 42:31] Bringing in a technical analytics engineer to the business team[42:31 - 45:33] The knowledge gap and bridging it with a framework[45:33 - 49:39] Alignment and building the semantic layer[49:43 - 56:36] Lightning Round[56:38 - 01:01:30] Takeaways[01:01:32 - 01:05:50] Three Questions

Takeaways with Gabi Steele and Leah Weiss

AI: No one wants your models. w/ Andrew Eye

You can’t buy a predictive model off the shelf. And if you COULD… would you? It’s the old “build versus buy” debate and with data in a constant state of change, building and deploying AI is more challenging than ever.Join Tim, Juan, and Andrew Eye, CEO at ClosedLoop.ai, to discuss the challenges of AI deployment, AIOps, and maintaining models as data changes.Key Takeaways[00:08 - 01:45] Cheers for AI[01:48 - 04:44] Taking advantage of incoming data and considering the data footprint[04:45 - 05:55] Building the most accurate and useful model, no best solution[06:01 - 09:26] AI strategy and developing prediction models[09:33 - 15:17] Should you build or should you buy?[15:21 - 20:02] Verticalized AI versus Horizontal data science and AI-oriented solutions[20:04 - 23:16] Build a custom predictive model[23:42 - 25:24] The concept of "feature drift" in AI[25:26 - 30:02] Vertical AI and the opportunity with healthcare[30:07 - 33:52] The concept of a click and the idea of semantic contracts[33:53 - 38:08] Great minds designing AI for Facebook instead of healthcare[38:12 - 39:41] Designing AI for good is addictive[39:47 - 41:27] The next generation and the promise of AI that can really impact lives[41:41 - 47:57] Lightning Round[48:02 - 52:03] Takeaways[52:05 - 53:24] Three questions

Takeaways with Andrew Eye

Knowledge: the missing piece to understanding your business

The most effective way to understand your business is the special balance between things like tribal knowledge and extracted knowledge, technical teams and business teams, confidence and skepticism.Join Tim, Juan, and Loris Marini, CEO of Discovering Data, as they zoom in to find that balance and the humility it may take.Key Takeaways[00:11 - 03:10] Introduction & Toasts[03:13 - 05:32] What's the weirdest place you've ever misplaced your keys?[05:44 - 08:01] The most effective way to follow the business is to follow the money[08:02 - 10:35] Thinking about business literacy and how to learn that perspective[10:37 - 12:35] How do we make change happen?[12:37 - 13:39] Having genuine conversations with people[13:44 - 18:19] Coaching, psychological safety, and communicating better at work[18:20 - 22:10] The definition of creating knowledge within an organization[22:11 - 25:34] Knowledge, insight, and levels of knowledge distributed in your team[25:35 - 28:52] You've got to spend time and you have to have a platform[29:10 - 31:48] We can do better than "just the smartest will survive"[31:57 - 35:56] Incentives are about subtle things like a team that listens, not just pay and flexible time[35:59 - 39:35] A data therapist and becoming more aware of issues[39:43 - 41:32] Advice to engineers who write the data pipeline, engineer the SQL queries[42:06 - 45:24] Lightning Round[45:28 - 51:10] Takeaways[52:06 - 53:57] Three questions

Takeaways with Loris Marini

Data empathy; you either got it or you don’t w/ Laura Ellis from Rapid7

Data needs to be in the hands of the people who really need it. Sounds simple, right? So where is the disconnect? If your teams aren’t partnering to get projects done, it’s often due to a lack of understanding of each others pain points. That’s where empathy is a must. Join Tim, Juan and special guest Laura Ellis from Rapid7 as they discuss supporting all the cooks in the kitchen when it comes to data projects, why that’s a hill to die on and HOW to actually get that done. Key Takeaways: [00:06 - 02:09] Intros & Toasts[02:09 - 03:49] What is the most controversial hill you'll die on?[03:52 - 05:30] Getting data into the hands of people who need it[05:30 - 07:01] Biggest problems for collaborating and making data easier to work with across the organization[07:05 - 09:32] Take a business problem, break it into a data problem[09:32 - 11:00] Breaking down problems, making it less intimidating[11:00 - 12:34] A user experience problem around data[12:30 - 13:52] Who is responsible for figuring out how the pieces of the puzzle fit together[13:55 - 17:22] Understanding centralization, decentralization, and embracing ownership[17:25 - 20:48] Identifying issues and opportunities research provides for enabling data access[20:48 - 22:31] Internal user research and Rapid7's "data therapist"[22:31 - 24:33] Business literacy and data literacy, cross-functional learning[24:28 - 26:39] More technology and more tools[26:37 - 33:14] Putting data into the hands of people who need it, making money and saving money[33:20 - 36:36] Gathering community and team content from passionate people[36:34 - 38:24] Brainstorming definitions of success[38:24 - 41:31] Start talking to people, do surveys, understanding the monetary impact[41:39 - 43:31] Dive into the details[43:36 - 47:23] Lightning Round[47:27 - 54:09] Takeaways[54:10 - 57:57] Three questions