Dave on Data

Have you heard of a confusion matrix? It's a powerful tool for understanding the quality of your machine learning models. Got 2 minutes for today's ML mini-lesson? ?? P.S. - If you're ready to build machine learning skills, I want to help. I've got 10 hours of free crash courses covering: Python Logistic regression K-means cluster analysis Decision tree machine learning **AND** Weekly machine learning tutorials. Join 7,715 DIY data scientists building skills: https://lnkd.in/e7fVrjxC

Over the years, I've been paid to write code with: C++ Visual Basic 6.0 C# Java R Python The only constant? SQL Even in 2025, you can't go wrong learning SQL.

How do you know if your random forest model is awesome? You check the model's oob predictive performance! Today's machine learning mini-lesson is 3 minutes. ??

Spoke to a client today. Their entire organization has access to Python in Excel. 2025 is the year of DIY data science.

7,715 DIY data scientists will receive a new machine learning tutorial tomorrow morning. It's the second in a series that is teaching these professionals a real-world skill: Hierarchical clustering with Python. But that's not all. These professionals are also building skills with 17 hours of free crash courses on: SQL Python Logistic regression K-means cluster analysis Decision tree machine learning Start building skills this weekend: https://lnkd.in/e7fVrjxC

It takes only 3 lines of Python code to train a random forest. The "how" of machine learning code is the easy part. The "why" of the code is what matters. ??

Python in Excel can be your gateway to DIY data science in 2025. Here's how: 1 - Learn the subset of Python you need for analytics. 2 - Learn k-means clustering. 3 - Learn decision tree machine learning. 4 - Apply to a dataset at work. Here's an example: A - Clustering users based on behaviors. B - Using a decision tree model to analyze the clusters.

So many organizations still analyze free-form text data by hand. STOP! The process is costly, biased, and produces dubious insights. This is one area (among many) where the LinkedIn Feed distorts reality. It's still very common for organizations to do the following: 1 - Capture text data from some process (e.g., surveys, chats, etc.). 2 - Put the text data in a Microsoft Excel workbook. 3 - Have a human periodically read the text, interpret it, and then label it. 4 - The labels are then used to make business decisions. 5 - Periodically, this onerous task is reassigned to another human. 6 - Every human interprets the text differently. Here's the good news. Analyzing text data in a standardized way is possible using Python. The attached PDF explains one of the most important concepts in text analytics - tokenization. You can think of tokenization as the first step in making free-form text understandable to the computer. If you like what you see in the PDF, I want to let you know that I will be teaching a full-day course on Text Analytics this May. The course is part of a 3-day live virtual ML bootcamp - no travel required! Everything you need is included: Live training where I answer your questions. 4-hour Python Quick Start online tutorial. 15 hands-on labs to build skills. 1-on-1 help if you get stuck. Learn more here: https://lnkd.in/g7FsWEPd BTW - Check with your manager about reimbursement. TDWI is an approved training vendor with many organizations. Stay healthy and happy data sleuthing! ?? BONUS ALERT If you attend all three days of my hands-on machine learning courses, you will receive my "Introduction to Machine Learning" online course as a bonus. That's a $599 value!

Today's machine learning mini-lesson is all about Python code. When using scikit-learn, you often need to encode your labels to train an ML model. Got 2 minutes to learn how? ?? P.S. - If you're ready to build DIY data science skills, I want to help. Each week, I send an ML tutorial to 7,715 professionals. But that's not all. These professionals are also learning with my free crash courses: SQL Python Logistic regression K-means cluster analysis Decision tree machine learning Get started today: https://lnkd.in/e7fVrjxC

? Less Than a Week Away! Are You Missing Out? Dimensional Modeling Meets Machine Learning – and it’s more relevant than ever. Join David Langer founder of Dave on Data & Patrick O'Halloran of WhereScape as they uncover: ? Why dimensional modeling is still a game-changer for ML & data science ? How automation speeds up model design & delivery ? Real-world use cases proving its impact ?? Tuesday, March 18 | ? 10 AM MDT | Noon EDT ?? Secure your spot now: https://ow.ly/xcll50VewIt #MachineLearning #DimensionalModeling #DataScience #ETL #DataWarehouseAutomation #WhereScape