Human37’s Post

🔍 Did You Know? KNN doesn’t build a parametric model of the data. Instead, it classifies new data points by comparing them to the k nearest points in the training set. #MachineLearning #DataScience #KNN #TechFacts

Q: Why did you adopt a dataset for #LoveData24? A: "I enjoy the annual tradition of exploring a new dataset! I learn something new and get to explore ICPSR's vast treasure trove of data." You can adopt a dataset, too! Visit https://lnkd.in/ei8YR4q

The new HDX Humanitarian API (HAPI) allows you to easily query 44 key humanitarian indicators to get disaggregated data in JSON and CSV formats. Here's a great example of the sort of interactive visualizations that can be built with it! We'd love to see what you build with HDX HAPI. Learn more about HAPI: https://lnkd.in/d65F7suC See more of this example: https://lnkd.in/dbmUAQ9M OCHA Centre for Humanitarian Data

🚄 Speed up your data workflows! Get a taste of the powerful PyKX in this analytics comparison to Pandas. In this short example, I show PyKX's ability to run analytics (such as correlation) significantly faster than Pandas. Learn more about PyKX here: https://lnkd.in/g5Qajy4H

📦 {ucimlrepo} v0.2.0 is out! 🎉 🚀 Faster data importation for large data sets 🛠️ Improved error handling 🐛 Fixed whitespace issues in variable names Get easy access to UCI Machine Learning Repository datasets in R! Repository: https://lnkd.in/gxVV3AKu CRAN: https://lnkd.in/gEpeUfdv #rstats #DataScience

🔍 Curious about the potential of LLMs in interpreting tabular data? Our latest blog post explores the effectiveness of various LLMs in analyzing tabular data through different prompting strategies and data formats. We delve into the comparison of text-based and image-based table representations, shedding light on the impact of representation and prompting on LLM performance. For a deeper understanding, check out the full article here: https://bit.ly/42MNTeI #DataAnalysis #LLMs #TabularData 📊

🔎 When it comes to data, you really do want to become a missing explorer. The naniar package in R (by the inimitable Nicholas Tierney) gives you "principled, tidy ways to summarise, visualise, and manipulate missing data." Quickly make summary tables and charts of missingness, create shadow matrices, see missing co-occurrence in UpSet plots, and much more. 👻 Don't be haunted by missing data. Check out naniar: https://lnkd.in/dQiJycTF

Live from dbt Labs #Coalesce! Don't miss our session tomorrow at 1:15 PM (PST), room Jasmine A, where Connor Carreras will dive into optimizing data workflows, boosting query performance, and enabling seamless data transformations with Firebolt and #dbt. Book time with our team on-site to discuss your data challenges: https://lnkd.in/dFSYXKQX #DataTransformation #DataEngineering

Tweeq Data Platform: Journey and Lessons Learned: Clickhouse, dbt, Dagster, and Superset https://lnkd.in/euK9xUG5 <-- one thing that struck me is how ubiquitous CDC is now becoming in these kinds of diagrams. A good direction for the industry.

Hello Text-to-SQL. I have a little-big ask for any CTOs and data folk out here from a colleague of mine. Martin Hao has completed building an MVP of a text-to-SQL tool and is looking for product feedback from CTOs or Head of Data folk. Here is a 90sec video tour of what's in the works: https://lnkd.in/g33GbGNW Please reach out to Martin Hao directly who is eager to learn how to improve this product with your input / feedback / questions / suggestions. Or share with people who you think might be interested in a tool like this. Thanks a thousand. ps. SELECT product_feedback FROM Customers WHERE role LIKE '*Data*'; :) #data #sql