Machine Learning and Artificial Intelligence in Power BI

Download and install Power BI Desktop from the official page, open pbix reports, explore the start panel, data window, and model window, and build a linked data model.

Import your dataset into Power BI from flat files, databases, online services, and web sources; use the query editor and M code to transform data and load a CSV dataset.

Explore data quality with visual tools in Power BI, using column quality, column distribution, and column profile to assess valid and erroneous values, duplicates, and statistics across datasets.

Explore Power BI data pre-processing with transformations across home, transform, add columns, view, and help tabs, and perform data cleaning, sorting, deduplication, filtering, and data-type checks.

Configure and teach the visual Q&A in Power BI, manage synonyms and term mappings, review user questions, and curate suggested questions to enable natural language data queries.

Configure the Q&A visual in Power BI by building synonyms, reviewing questions, and teaching the model to understand natural language queries for metrics like total sales.

Identify top regions by back order and analyze demand types in Power BI. Propose questions like top ten plants by back order and save the resulting static chart.

Explore the key influencers chart in Power BI and learn how it automatically analyzes data, identifies key influential elements, and explains sales changes with an average line.

Explore the key influencer chart's main segments to see how six segments, with averages and case counts, reveal which factors drive total sales and profit in Power BI.

Understand that correlation does not imply causality; correlations may result from coincidence or a third factor. The cancer versus mobile phones example shows patterns do not prove cause and effect.

Analyze back order percentages using a hierarchical schema in Power BI across plant, month, region, and demand type, highlighting outliers with AI insights such as plant 0288 and June.

Explore Power BI's time series forecasting that uses exponential smoothing to predict future values on line charts, with configurable forecast length, confidence intervals, and seasonality.

Create a time-series line graph with anomaly detection in Power BI, remove forecasts, add explanation variables, and analyze 2015 anomalies by linking related tables.

Learn to install Python with Anaconda, create and activate a virtual environment for Power BI, install libraries, and link the environment to Power BI.

Install the Pi Card auto machine learning library in Power BI using Python and Anaconda prompt, activate its Python 3.7 environment, then pip install PyCaret 2.0 to train models.

Learn to run Python scripts in Power BI to import, preprocess, and visualize data. Follow get data, transform, and visualizations steps to insert code and use dataset variables for charts.

Learn seaborn basics for creating and customizing statistical plots with matplotlib underneath, using a data frame to generate lm plots, scatter, box, violin, kde, and more.

Select the right chart type based on data types and the desired representation. Use scatter plots for relationships, heatmaps for correlations, and bar, swarm, cd and ds plots for distributions.

Learn to preprocess health insurance data in Jupyter notebook using Python, encode categorical variables with getdummies, discretize age into young, adults, seniors, and create body mass categories for Power BI.

Integrate machine learning and artificial intelligence in Power BI with open source tools and Python libraries, training models with a few lines of code and AI, ML concepts.

Explore supervised learning with classification and regression, unsupervised learning including clustering and dimensionality reduction, and reinforcement and deep learning, with practical examples like spam detection and market basket associations.

Describe the phases of training machine learning models in Power BI, including data collection, cleaning and pre-processing, feature engineering, model training with hyper parameters, evaluation, cross-validation, and prediction.

Explore common machine learning algorithms across supervised, unsupervised, and regression tasks in Power BI, including SVM, k-NN, neural networks, linear regression, SVR, decision trees, random forests, and clustering.

Deploy models in Power BI by either training inside Power BI or externally and loading a binary model for predictions.

Learn to obtain and interpret regression evaluation metrics in Power BI using sklearn, including mean absolute error and RMSE, to assess model performance.

Explore AutoML to automate data preprocessing, model selection, and hyperparameter tuning, then assemble and deploy models in production using PyCaret for rapid Power BI analytics.

Train and optimize binary classification models with Pycaret on the credit dataset, using setup and preprocessing, compare models, and tune hyperparameters with cross-validation.

Evaluate Pycaret models using plot_model to inspect AUC and precision-recall, review feature importance and confusion matrices, and validate performance on unseen data before finalizing, saving, and deploying the model.

Explore regression with Pycaret in Power BI, comparing models from linear to gradient boosting and Catboost, and evaluate performance using residuals, r-squared, RMSE, and validation curves.

Learn how to add model evaluation graphs to Power BI dashboards by importing image files, creating dynamic queries, and using the simple image visual to display evaluation metrics.

Launch a Jupyter notebook, import diamond dataset via PyCaret, and apply normalization, transformations, and multicollinearity handling before training Catboost regressor and light gradient boosting machine, predicting prices in Power BI.