Scott Czepiel

One of my favorite ML YouTubers is Joshua Starmer PhD. One thing he commented offhanded (or not,maybe he rehearsed that sentence 25 weeks, who knows), and I am totally paraphrasing here is that Stats is not just Data Science, and Data Science is not just Stats. Remember the formulas I was playing around with? The simple, updated formulas on Standard Deviation and Mean. Today, I put them to the test with simulated datasets mimicking real-world data additions. and the results, wow. This not only saves computing time, which adds up quick, but it also saves computing power. #DataScience #MachineLearning #DataAnalytics #IDoNotKnowWhatIAmDoing

14

🌎 👀 "GeoSynth" >> created with Fused App Builder. Not just to show off, but to demonstrate how easy it is, for someone who just started playing with Streamlit 3 weeks ago - you should try it out! Solving some real painpoints for geospatial data scientists -- in multiple stages of the ML dev pipeline: 1. 👷‍♀️Data Curation -- building ML training datasets quickly, fusing multiple resources from multispectral satellite data to elevation models. These can then be streamed to Hugging Face for use by the GIS community! 2. 🧐Model Inspection and Eval -- Comparing ground truth (Cropland Data Layer - CDL) with the output of a Transformer model i built 🛠️(on top of 10,000 data cubes curated earlier via the same app). 📈 Chosen metric is IOU for each crop type predicted, spot-evaluated at the central map tile! #ml #geospatial #datascience >>> I demoed Geosynth at yhe Fused Linkedin webinar- https://lnkd.in/gT6AQxkr

112

17 Comments

Mintel has launched an AI to help you tap into our massive depth of data and market reports in seconds. I am truly proud of Mintel and extremely excited about this application of AI. The hype around LLMs such as Chat GPT has been incredible this past year. And for once, I think the hype is actually well founded. I'm proud of how quickly Mintel has responded to the general availability of this technology. Mintel Leap is an actually useful and trustworthy implementation of generative AI. While generative AI has been shown to produce some very interesting content, I have long been an advocate that this technology should augment human intelligence, not replace it. Mintel Leap answers your market questions by citing Mintel's deep research, data and human-generated market reports. It helps you synthesize our data and get you to the insight faster. Like, super fast. I'm excited for how this is going to impact our customers. Customers that have long trusted our data and research and who can now shave days/weeks from their processes.

97

3 Comments

In my experience as a data leader at companies of all scales (Target, Ask.com, Loom, Shipt & GoFundMe), these are the most common challenges with data analytics. That’s why I built Zetta. What’s your biggest challenge with data analytics?

41

6 Comments

New evaluation benchmark research from Databricks Mosaic Research: "There is no single best model or paradigm. From open-source options to retrieval strategies, different solutions excel in different scenarios." Other interesting observations: - Small models were surprisingly strong performers: they mostly performed similarly to their larger counterparts, and sometimes significantly outperformed them - Open-source models are off-the-shelf viable for at least two high-frequency enterprise use cases (Text2Json and function calling) Congrats to Pallavi Koppol, Erica Ji Yuen, Kartik Sreenivasan, Yue (Andy) Zhang, Sam Havens, Michael Carbin, Jonathan Frankle, and Matei Zaharia. Incredibly valuable work. Full blog post here: https://lnkd.in/g9_CRiHD

23

1 Comment

Graphy now powers over 100,000 data storytellers 📈 This quote from Stan Massueras really nails it: "Great ideas are only impactful when communicated clearly. Graphy enabled us to transform data into compelling stories, and those stories into meaningful action." Thanks to everyone using Graphy to turn data into stories that drive decisions, inspire teams, and make complex insights easy to act on ❤️

54

10 Comments

If you're looking to build internal data apps, but can't yet stomach the recurring SaaS cost - especially relevant for data-centric early-stage startups... Would highly recommend taking a look at https://streamlit.io/ - a Python framework for data apps 🐍 Can deploy on GCP Cloud Run, or your preferred cloud platform, and away you go. Clean interactive front end, using Python only. For larger teams, I would definitely be looking at the likes of Count (data canvas pioneers) and Hex - but at this stage, Streamlit on Cloud Run is working very nicely 👌 #DataScience #DataEngineering #DataOps #Startups

15

𝗦𝘁𝗿𝗲𝘀𝘀𝗟𝗲𝗻𝘀 is now an official product! The AI tool can monitor public figures such as the folks on the Fed, politicians, and C-suite execs at public companies, and model them individually. If their brain is in front of their mouth or they're potentially obfuscating, StressLens can tell. Check it out here: https://lnkd.in/eBxWJBPu Stay tuned for more updates on StressLens and its analysis of public figures like Biden and Trump.

122

19 Comments

There is a powerful and transformative change that is happening right now in the world of business insights creation and activation that can deliver AI readiness that is accurate, secure, fast, and scalable. I’d recommend starting with this video if you’re on point to make AI real for your organization.

10

Excellent study from Damian Rollison and the SOCi, Inc. team which seems to confirm my longtime hypothesis around citations (https://lnkd.in/gxZQwdx3): the only citations that are valuable are the ones that rank in Google for the keywords you want to be known for. https://lnkd.in/g49DMSeN

38

11 Comments

Huge win on #Quarterly Earnings for Palantir Technologies (#PLTR +13.26% pre-market) If you were on the #Earnings #Call, you might have heard Shyam Sankar talk about their core product: #Data #Ontology. The CTO describes how Ontology gives Palantir an unprecedented 10+ year advantage in the #AI space, not for competing with the #ChatGPTs models of the world, but by actually providing enterprise framework/software to deploy 3rd party AI models beyond your design, content writing and #meme producing teams. We are talking about creating your JARVIS from Iron Man; scaleable to every end user in your company - with access points, seamless resource sharing, and actionable analytics with tailored cross-department recommendations. And it all starts with Ontology. We've written about Data Ontology and how this can be deployed in a Marketing environment. Granted this is child's play compared to PLTR's Apollo, Foundry, and Gotham offerings; yet this is a Day 1 thinking approach to a complex and ever-evolving endeavor. Let me know what you think: https://lnkd.in/evtcWeTb

2

First rule of trying to hire a "unicorn" - Don't. Literal unicorns don't exist nor does the mythical person you're trying to find, so why have a plan that is built upon fantasy? Instead, treat the desire to find a unicorn as a sign that you need to rethink your role definition. Whether it be chunking the role into multiple roles (if you have the budget to do so) or more ruthlessly prioritizing your job requirements, set you and your future hire up for success by having expectations that can exist in an individual. Arguing with reality never ends well. The sooner you abandon the unicorn hunt, the sooner you make the best possible hire and accelerate progress towards your goal - even if it's not quite the level of acceleration you dream of, it's still faster than you're going today.

135

21 Comments

How do you prove the value of your data and AI investment? I love the “data tax” concept. Everything you do will have a data element whether it be an hour of your data analysts time to understand the performance of your latest marketing campaign, the need for data engineers to build and then maintain pipelines for a new data source, implementing more efficient processes using AI, right through to the more obvious license costs for your big ticket tech. These activities are often considered central or core services and as a result are just factored in as a cost of doing business. This makes seeking investment for new resource, projects, tech and tools really challenging. If you take the time to look at all your initiatives and identify the “data tax” involved you can start to really demonstrate the ROI of your data and AI effort. This type of language also resonates well with your FD and C-Suite colleagues, but more importantly provides the foundations of a great story you can share across the business as part of your data literacy journey. Has anyone else used this concept as a way to prove the value of your data and AI effort?

13

4 Comments

Select AI automates the creation and population of your #vector database using your raw data, helping you to get started with Retrieval-Augmented Generation. https://lnkd.in/ex2YiU9j

4

Often we report point predictions and accuracy from that when we really should be talking confidence intervals. https://lnkd.in/e5aeAfxH

8

2 Comments

My team at Spotify builds products for data practitioners to discover, analyze, and share insights. Data practitioners are often grouped into a single job title: “Data Scientist”, but this title doesn’t capture the diversity of skills and work these users are actually doing. To create products that best fit the needs of our users, my amazing teammate and researcher, Serena Fang, and I built a framework to better understand who our users really are: ✨ the Data Science Personas ✨ Our blog post covers the mixed-methods approach we took, our learnings and who those 6 personas are. If you work in data you can likely identify with at least one of the personas and I’d love to know: which persona are you?

61

4 Comments

What does marketing in a post-cookies world look like? https://lnkd.in/gN7BPfjQ #datascientist #dataanalyst #cfo #cmo #productmanagers #founders #cofounders #cxo #ceo #cpo #bizops #revops #operations #b2bsaas #marketing #marketingmanagers

1

Data Canvas is new AI powered feature in preview on GCP that simplify the data analysis of BigQuery stored data, the demo look really promising: https://lnkd.in/em2JM8cM

1