Microsoft Copilot for Data Evaluation and Data Engineering

Profile a real-world csv dataset and clean it with Copilot to flag anomalies, fix missing values and invalid formats, and produce a ready-to-analyze dataset.

Imputing missing values, fixing date formats, and adding calculated columns with Copilot; explore mean, median, mode, group-based and predictive imputation, and create the customer age group and first/last name columns.

Upload a file to Copilot to analyze a sales orders dataset, perform data quality checks and exploratory data analysis, and outline a Python-based data pipeline for cleaning and validation.

Use Copilot to generate a complete Python data quality pipeline that cleans a data set, logs actions, normalizes column names, strips whitespace, drops duplicates, and validates data formats and logic.

Leverage copilot to translate code across languages and generate data quality visuals with matplotlib and seaborn, then build a data pipeline with visuals that reveal outliers, multi-country customers, and trends.

Explore synthetic data, artificially generated data that mimics real data's statistical properties while stripping out personal information, enabling privacy-preserving AI training, testing, and analytics.

Generate synthetic data with Copilot by building a date table with year, month, date, day of week, and ISO week, then export a CSV via Python.

Explore the risks of synthetic data, including data quality and accuracy, privacy leakage, data drift, overfitting and underfitting, and bias amplification that can lead to unfair or discriminatory outcomes.

Examine risks of synthetic data, including edge-case underrepresentation, minority patterns, validation challenges, and the one-way road that prevents reverse engineering to original data, in fraud detection and safety-critical settings.

Document the purpose and choose a generation strategy (statistical methods, GANs, VAEs, or rule-based). Ensure realism with KS tests and PCA visuals, apply privacy safeguards, and separate data.

Explore how the M365 Copilot architecture grounds prompts by using the Microsoft Graph API to access authorized data, then queries the Azure OpenAI service to return contextual responses.

Explore the M365 Copilot interface for organizations, including licensing, admin controls, work and web views, and how Copilot analyzes cloud data from OneDrive and SharePoint with enterprise data protection.

Explore M365 Copilot agents, especially the analyst agent, to analyze data, compute correlations, visualize with a heatmap, and forecast sales using exponential smoothing.

Discover Copilot inside Excel, using natural language to analyze data, create pivot tables and charts, apply Python analytics, and leverage autosave and prompts gallery.

Sort and filter data with Copilot on an employee and department dataset, applying a custom sort by bonus percentage and multi-criteria filters to identify promotable, high potential, outstanding performers.

Learn to generate custom Excel formulas with Microsoft Copilot, creating active status and age columns from termination dates and date of birth, and split full names and emails using delimiters.

Create calculated columns with Copilot to determine years of service using hire and termination dates, and calculate bonuses from base salary and bonus percentage, using date diff and if logic.

Create pivot tables in Excel with Copilot by formatting data as a table and generating tailored analyses using department, job title, active status, work location, and the average base salary.

Harness python with Copilot in Excel to analyze data, generate insights, and build visuals from a pandas data frame. Explore statistics, trend analysis, and ML model training on HR data.

Explore microsoft fabric, an enterprise analytics platform that unifies data ingestion, lakehouses, data warehouses, real-time intelligence, Power BI dashboards, governance, and Copilot, all within one organization-wide data lake.

Navigate the Microsoft Fabric interface, create a workspace with lakehouse, SQL analytics endpoint, notebooks, data flows, and explore One Lake data catalog and Copilot’s role in data evaluation.

Explore the ETL flow in fabric, from data extraction to loading into a storage solution. Create and use semantic models for Power BI reports and dashboards.

Leverage Copilot in Data Flow Gen2 to create ETL pipelines in the data fabric, performing source-to-destination transformations with drag-and-drop steps and natural language guidance.

Showcases copilot in dataflow gen 2 for renaming fields, joining datasets with left joins, merging steps, and loading a curated table to a SQL warehouse.

Deploy a SQL database on Microsoft Fabric, use Copilot for databases to write, explain, and fix SQL queries, and identify top ten highest paid employees by base salary.

Explore how Copilot generates SQL queries and helps diagnose non-numeric data issues, such as nvarchar fields, while guiding you through fixing query errors in a practical data evaluation workflow.

Explore data lakehouses and warehouses in fabric, create a lakehouse and a warehouse, run Copilot-generated SQL queries, and build semantic models for Power BI to analyze trends with notebooks.

Use notebooks in the fabric environment with Copilot to load data from a lakehouse into a pandas data frame, cleanse with data wrangler, and explore ml-ready data preparation.

Clean data in notebooks with data wrangler and natural language. Leverage Copilot and AI tools in notebooks for machine learning and auto machine learning workflows.