For Personal Data Analysis!
- Power BI and Microsoft Teams are enabling you to view and analyze your personal Microsoft Teams activity in 1-click with a new Power BI ‘Teams activity analytics’ report – now in preview. [Microsoft]
For Story Tellers!
- Are you sending a screenshot of your report to your stakeholders? Try Smart Narrative visual. It can point out trends, understand key points and dynamically change when you cross-filter. [Microsoft]
- A sudden spike in sales, change in trend, change in temperature is difficult to discover in huge volume of data. Try Anomaly detection in line chart. [Microsoft]
For Performance Tuning Warriors!
- Is your data import in Power BI desktop taking too long? If yes this could be due to number of evaluation containers. Power BI Desktop leverages multi-threading technology to optimize query performance when importing data or when using DirectQuery. You can control this behavior and their by influence the level of parallelism in PowerQuery. [Chris Webb]
- Download best practice rules and run these rules in Tabular editor to identify and improve model performance [Microsoft]
For Governance and Deployment Engineers!
- PREVIEW: Power Platform admin center is upgraded to view and manage Power BI cloud and on-premises data sources and gateway clusters. [Microsoft]
- LIVE: Microsoft identified a bug recently which applies only to Azure Logic Apps. Using the SAP connector for Azure Logic Apps with the first release of June gateway, may cause intermittent connectivity failures or may lead to corrupted/malformed data to be returned . June gateway release 2 with version “3000.86.4” addresses this problem. [Microsoft]
Behind the scenes! Not of light hearted
- Power BI anomaly detection maintains simplicity outside – it just expects time series data and sensitivity, and rest on selecting the features, best-fit algorithm, detection the outliners done inside. It applies spectral residual on time series data, then applies Convolutional Neural Network (CNN) to train the model. Know more about this algorithm [Microsoft]
- When you pick KPI, say house price or rating, to analyze, the Key Influencers visualization uses machine learning algorithms provided by ML.NET to figure out what matters the most in driving metrics. For numerical metrics, such as house price, it runs linear regression and uses SDCA regression. For categorical metrics, such as rating, it runs logistic regression. Read more about this here [Microsoft]