Skip to main content

November 2026

Every post published in November 2026, newest first. Pick a day to jump straight to it.

Posts
802
Topics
75
Reading time
91 h
Active days
15

Posts 91–120 of 802 · November 13–18, 2026

Page 4 of 27

Wednesday, November 18, 2026

4 of 30 posts

  1. How to Use warm_start to Add Trees to a Random Forest in scikit-learn

    Grow a random forest in stages with warm_start=True: raise n_estimators, call fit again, and scikit-learn adds only the new trees, so you can score the forest as it grows. Real output from scikit-learn 1.9.1, including the warning and the error you get when n_estimators doesn't go up.

  2. How to Split Data into Training, Validation and Test Sets with scikit-learn

    Split a dataset three ways with two calls to scikit-learn's train_test_split: a test set kept for the very end, a validation set for checking the model while you tune it, and a training set. Real output for a 60/20/20 split of the California Housing data with scikit-learn 1.9.1.

  3. Fixed: "No charts in this section" on the MLflow System Metrics Tab

    A short MLflow run showed System metrics (0) and "No charts in this section", although system metrics were on and longer runs had charts. MLflow logs a sample only after waiting one sampling interval and finding the run still going, so a run shorter than the interval records nothing. Sample more often than your shortest run lasts.

  4. Fixed: scikit-learn Random Forest Predictions Change Slightly Between Calls with n_jobs=-1

    np.array_equal said two random forests with identical trees predicted different values, and even one forest predicted different values on two calls. With n_jobs=-1, scikit-learn adds the trees' predictions in whatever order its threads finish, and the rounding changes the 15th decimal place. Compare with np.allclose, or predict with n_jobs=1.

Saturday, November 14, 2026

11 posts

  1. Build an Overtime Timesheet App in Power Apps — Part 10: Build the Departments Screen

    Build the screens where admins manage departments in the overtime app: a list with each department's approvers and status, and one screen that adds or edits a department, with a people picker, an Active toggle and a check for duplicate names.

  2. How to Save a Toggle to a SharePoint Yes/No Column in Power Apps

    Show a SharePoint Yes/No column in a Power Apps toggle and save it back with Patch: the toggle's Default for new and existing items, Checked in the Patch record, and what SharePoint shows afterwards.

  3. How to Add a Toggle in Power Apps

    Add the modern Toggle control to a Power Apps canvas app: insert the right one, set its label and starting value, and read whether it's on with Checked in your formulas.

  4. How to Save People to a SharePoint Person Column with Patch in Power Apps

    Save the people chosen in a Power Apps people picker to a SharePoint Person column: put the combo box's SelectedItems in the Patch record, and the column holds exactly the people in the box.

  5. How to Show Saved People in a Combo Box in Power Apps

    Show the people already saved in a SharePoint Person column when an item opens in Power Apps: set the combo box's DefaultSelectedItems to the column, reset it as the screen opens, and keep them selected with AllowExternalSelectedItems.

  6. How to Add a People Picker in Power Apps

    Turn a modern combo box into a people picker in Power Apps: Choices on a SharePoint Person column, filtered by what's typed, so the start of a first name or surname finds people the column can save.

  7. How to Choose When a Text Input Updates in Power Apps with TriggerOutput

    Set a modern text input's TriggerOutput in Power Apps: Keypress updates Text on every key, FocusOut when the box loses focus, and Delayed after a short pause, which suits checks that ask a data source.

  8. How to Show the Names in a SharePoint Person Column in Power Apps

    Show the people in a SharePoint Person column as one line of names in a Power Apps gallery: Concat over the column's records with DisplayName, for a column that allows several people.

  9. How to See the Code of a Control in Power Apps Studio

    Open View code in Power Apps Studio to read a control's code: its type and version and every property you've changed, in one place. Useful for checking what a copied control kept.

  10. Fixed: "The First Argument to the GroupBy Function Cannot Be Blank" in Power Apps

    A Power Apps screen showed the error banner "The first argument to the GroupBy function cannot be blank" as it opened, because a label grouped a context variable that the screen's first load hadn't set yet. Testing the variable with IsBlank before grouping removed the banner.

  11. Fixed: Delegation Warning. The "Trim" Part of This Formula Might Not Work Correctly on Large Data Sets

    A Power Apps duplicate-name check that filtered a SharePoint list on Department = Trim(Name) got two delegation warnings, one for the Filter part and one for the Trim part. Trimming the name first with With and comparing the column with that value removed both.

Friday, November 13, 2026

15 of 34 posts

  1. MLflow Tutorial for Machine Learning in Python — Part 1: Install MLflow and Track Your First Experiment

    Install MLflow 3.18 with uv, start your own tracking server on your computer, and log your first machine learning experiment from Python: parameters, metrics, tags and a chart for a scikit-learn house price model, then compare the runs in the MLflow UI.

  2. How to Keep MLflow Files Out of Git with .gitignore

    MLflow can create mlflow.db and an mlruns folder inside your project without asking, and a tracking server adds mlartifacts. Add three lines to .gitignore, prove they work with git check-ignore -v, and keep tracking data out of your commits. Real output from MLflow 3.18.0 and Git.

  3. How to See Where MLflow Stores Runs and Artifacts

    Find the two places an MLflow tracking server keeps your work: the mlflow.db SQLite database for experiments, runs, parameters, metrics and tags, and the mlartifacts folder for files. Follow one run from the UI to its files and rows, and read the database safely with Python's sqlite3. Real output from MLflow 3.18.0.

  4. How to Compare Runs in the MLflow UI

    Compare MLflow runs side by side: add metric and parameter columns, filter with queries such as metrics.r2 > 0.8, sort by a metric, chart the runs, and read the Compare page's parallel coordinates, scatter, box and contour plots and its diff-only tables. Real screenshots from MLflow 3.18.0.

  5. How to Log a Matplotlib Figure to MLflow

    Save Matplotlib charts straight into an MLflow run with mlflow.log_figure: no temporary files, any folder inside the run, PNG or SVG by file name, and the pyplot version with plt.close. See the logged chart in the run's Artifacts tab. Real output from MLflow 3.18.0.

  6. How to Plot Predicted vs Actual Values with Matplotlib

    Draw a predicted vs actual scatter plot for a regression model with Matplotlib's Figure class, save it as a PNG, and read what it tells you: where the model is right, where it misses, and why capped data makes a vertical line. Real chart from a scikit-learn random forest.

  7. How to Calculate MAE, RMSE and R² in scikit-learn

    Measure a regression model with scikit-learn's mean_absolute_error, root_mean_squared_error and r2_score: what each number means, how to read it in real units, and why comparing against an always-guess-the-average baseline makes R² easy to understand. Real output from scikit-learn 1.9.1.

  8. How to Train a Random Forest Regression Model with scikit-learn

    Train a random forest regression model with scikit-learn's RandomForestRegressor: what n_estimators, max_depth, random_state and n_jobs do, how to make predictions, and how to see which features mattered most. Real output on the California Housing data.

  9. How to Split Data into Training and Test Sets with scikit-learn

    Split a dataset into training and test sets with scikit-learn's train_test_split: what test_size and random_state do, why the same seed always picks the same rows, and why a model must be scored on rows it never saw. Real output on the California Housing data.

  10. How to Load the California Housing Dataset in Python with scikit-learn

    Load scikit-learn's California Housing dataset as a pandas table with fetch_california_housing: what its 8 columns and target mean, where the download is cached, and the $500,000 cap you need to know about. Real output from scikit-learn 1.9.1.

  11. How to Add Tags to an MLflow Run

    Label MLflow runs with mlflow.set_tag and set_tags, overwrite a tag, write a run description with mlflow.note.content, and see the system tags MLflow adds by itself, including the Git commit and branch. Real output from MLflow 3.18.0.

  12. How to Log Metrics in MLflow

    Record a run's results with mlflow.log_metric and mlflow.log_metrics, log a metric at many steps to get a curve, read the full history with get_metric_history, and see the error MLflow gives for a value that isn't a number. Real output from MLflow 3.18.0.

  13. How to Log Parameters in MLflow

    Record a run's settings with mlflow.log_param and mlflow.log_params: how MLflow stores every value as text, why a parameter can never change once logged (with the exact error), and the size limits to stay within. Real output from MLflow 3.18.0.

  14. How to Start, Name and End an MLflow Run

    Start MLflow runs with mlflow.start_run as a with block or by hand, give them names, check their status, and see what happens when your code fails inside a run: MLflow marks it FAILED. Real output and screenshots from MLflow 3.18.0.

  15. How to Create and Set an MLflow Experiment

    Group your MLflow runs into experiments with mlflow.set_experiment, which creates one when it's missing, or MlflowClient.create_experiment with tags. See the experiment's ID and artifact location, the error for a duplicate name, and where to find it all in the UI. Real output from MLflow 3.18.0.