Temperature and Top-p Explained: Why the Same Prompt Gives Different Answers
What the temperature and top-p settings in AI models actually do, with a worked numeric example, and which values to use for code, data…
Read article →AI & ML
Artificial intelligence explained by building it: machine learning, neural networks, LLMs, RAG and agents — plus how to build your own language model.
Using AI to Write Excel Formulas — and How to Check Its Work
Using AI to Write Excel Formulas — and How to Check Its Work
AI assistants can draft Excel formulas and VBA in seconds, but they also make confident mistakes. How to prompt them well and a 5-step…
Read article →7 Prompt Patterns That Get Better Answers from ChatGPT and Claude
7 Prompt Patterns That Get Better Answers from ChatGPT and Claude
Seven practical prompt patterns — context, examples, format, steps, constraints, delimiters and questions first — with before-and-after prompts you can reuse.
Read article →Linear Regression in Excel with LINEST, SLOPE and FORECAST — No Add-ins
Linear Regression in Excel with LINEST, SLOPE and FORECAST — No Add-ins
Machine learning starts with a straight line. Fit a regression in Excel with SLOPE, INTERCEPT, RSQ, LINEST and FORECAST.LINEAR, and learn how to read…
Read article →Linear Regression From Scratch in Python: How a Model Actually Learns
Linear Regression From Scratch in Python: How a Model Actually Learns
Code linear regression with gradient descent in plain NumPy, watch the model learn slope and intercept, then check it against NumPy and scikit-learn. The…
Read article →A Neural Network From Scratch in 30 Lines of NumPy
A Neural Network From Scratch in 30 Lines of NumPy
Build and train a two-layer neural network in plain NumPy to solve XOR — forward pass, backpropagation and gradient descent explained line by line.
Read article →Machine Learning Lesson 8: Project, Predict Late Invoice Payments
Machine Learning Lesson 8: Project, Predict Late Invoice Payments
Build a complete model that predicts which invoices will be paid late, test it, save it, and score new invoices into an Excel file…
Read article →Machine Learning Lesson 7: Pipelines, Cross-Validation and Avoiding Leakage
Machine Learning Lesson 7: Pipelines, Cross-Validation and Avoiding Leakage
Bundle cleaning and the model into one scikit-learn pipeline so nothing leaks from test data, and compare models fairly with 5-fold cross-validation.
Read article →Machine Learning Lesson 6: Clustering Customers with K-Means
Machine Learning Lesson 6: Clustering Customers with K-Means
Group shoppers into segments without any labels using K-Means. Scale features first, choose the number of groups with the elbow method, and describe each…
Read article →Machine Learning Lesson 5: Decision Trees and Random Forests
Machine Learning Lesson 5: Decision Trees and Random Forests
Decision trees learn readable if/then rules; unlimited trees overfit. Random forests combine hundreds of trees and show which features matter most.
Read article →Machine Learning Lesson 4: Classification (Which Customers Will Leave?)
Machine Learning Lesson 4: Classification (Which Customers Will Leave?)
Predict yes/no outcomes with logistic regression. Read accuracy, precision, recall and the confusion matrix, and choose your own probability cut-off.
Read article →Machine Learning Lesson 3: Train/Test Split and Honest Evaluation
Machine Learning Lesson 3: Train/Test Split and Honest Evaluation
Keep part of the data hidden from the model to test it fairly. Measure regression accuracy with MAE and R2 and always compare with…
Read article →