Task 3.3: Identify patterns and trends

• ARemove the Top N filter from the chart's filters

  Correct

  Explain the increase doesn't support Top N filters, so the chart must not use one.

• BSwitch the sales table to DirectQuery and keep the filter

  Incorrect

  DirectQuery isn't supported by the insights feature at all.

• CReplace the month axis of the chart with a categorical Month Name column

  Incorrect

  Categorical X-axis columns are unsupported unless they sort by a scalar column, so this doesn't help.

• DChange the chart to a line chart with the same filter

  Incorrect

  Explain the increase works on bar and line charts alike, and the Top N filter is still unsupported.

Analyze > Explain the increase or decrease looks for the factors behind a change in a chart. It doesn't support Top N filters, include or exclude filters, measure filters, non-numeric or filtered measures, Show value as, models with RLS or OLS, or DirectQuery and live connections.

• ARight-click Age, choose New group, and set the bin size to 10

  Correct

  Binning puts numeric values into equally sized groups, which appear as a new (bins) field.

• BRight-click the Age measure and set a bin size of 10

  Incorrect

  Age is a column in the Customer table, so there's no Age measure to bin.

• CSelect the ages in a chart, right-click and choose Group for each band

  Incorrect

  Grouping selected points by hand doesn't create equal-width bands for all ages.

• DAdd a percentile line at every tenth percentile

  Incorrect

  Percentile lines mark values on a chart; they don't count customers by age band.

Binning sets a bin size for numerical and time fields so values fall into equally sized groups. Right-click the field and choose New group, set the bin size, and a new field with (bins) appended appears.

• ADecomposition tree, with Rating as the value to analyze

  Incorrect

  A decomposition tree doesn't rank attributes by how much they raise the chance of a Low rating.

• BQ&A visual asking which customers rate low

  Incorrect

  Q&A returns visuals for questions; it doesn't analyze which factors influence the rating.

• CScatter chart of tenure against rating for each of the attributes

  Incorrect

  A scatter chart doesn't rank the attributes by their impact on the rating.

• DKey influencers, analyzing Rating by the attributes

  Correct

  It analyzes the data and ranks the factors that drive the selected value of the metric.

The key influencers visual analyzes a metric, placed in Analyze, against the fields in Explain by, and ranks the factors that matter. Explanatory factors must be defined at the table level of the analyzed metric.

• AAdd a trend line from the Analytics pane to the clustered column chart

  Incorrect

  A trend line fits the history; it doesn't project six future months with a confidence band.

• BChange the visual to a line chart, then set Forecast length and Confidence interval in the Analytics pane

  Correct

  Forecasting is available only for line charts with time data.

• CAdd error bars with a 95% confidence interval to the clustered column chart

  Incorrect

  Error bars don't extend the chart into future months.

• DAdd a Y-axis constant line on the visual at next year's forecast value

  Incorrect

  A constant line doesn't project future months or a confidence band.

Forecast predicts future values from historical trends and is only available for line charts with time data, where you can set inputs such as Forecast length and Confidence interval.

• AChange the visual to a clustered column chart without a legend

  Incorrect

  Anomaly detection is available only for line charts.

• BAdd Region to the line chart's Explain by field

  Incorrect

  Explain by fields don't make detection available; the legend still blocks it.

• CRemove the legend and show one region at a time

  Correct

  Anomaly detection isn't supported with legends, multiple values or secondary values.

• DRaise the sensitivity in the anomaly settings

  Incorrect

  No anomaly setting makes detection available while the chart has a legend.

The Anomalies feature in the Analytics pane identifies unexpected spikes or dips in time-series data on line charts. Anomaly detection isn't supported with legends, multiple values or secondary values in the line chart.

• ASelect the thumbs-up icon at the top of the window

  Incorrect

  The thumbs icons give feedback about the visual and the feature.

• BSelect an icon at the bottom of the explanation visual

  Incorrect

  Those icons switch how the insight is displayed, such as a scatter or ribbon chart.

• CSelect the + button at the top of the explanation visual

  Correct

  It adds the selected visual to the report as if you had created it yourself.

• DRight-click the point and choose Find where this distribution is different

  Incorrect

  That runs another kind of analysis; it doesn't add the current visual to the page.

Explain the increase or decrease opens a window with explanation visuals, a waterfall by default, which you can switch to a scatter, 100% stacked column or ribbon chart with the icons at the bottom. The + button at the top adds the selected visual to the report, where you can format it like any other visual; it's available only while you're editing the report.

• AValues predicted for future periods

  Incorrect

  That's what a forecast shows.

• BThe variability or uncertainty around each value

  Correct

  Error bars display variability or uncertainty in measurements.

• CThe average of all values on the chart

  Incorrect

  Error bars are drawn for each value, not as one line for the chart.

• DUnexpected spikes and dips over time

  Incorrect

  That's anomaly detection.

Error bars display the variability or uncertainty in measurements around each value. Forecasts predict future values, and anomaly detection flags unexpected spikes or dips.

• AIt's the highest or lowest value on the chart

  Incorrect

  Extreme values can still sit inside the expected range; being the largest isn't the test.

• BIt changed by more than 10% from the previous point

  Incorrect

  No fixed percentage applies; the test is whether the value leaves the expected range.

• CIt lies above the chart's average line

  Incorrect

  Anomalies are judged against the expected range, not against an average.

• DIts value falls outside the expected range that the algorithm computes

  Correct

  Power BI adds an expected range of values, and a value outside that boundary is marked as an anomaly.

Anomaly detection enriches a line chart that has time-series data with anomalies and the expected range of values. When a value goes outside the expected boundary, it's marked as an anomaly. You can format the anomalies and the expected range.

• AA plain text box with the number typed in

  Incorrect

  Typed numbers don't update with the data.

• BA card visual with the growth measure

  Incorrect

  A card shows the number, but not as part of your sentence.

• CA Q&A visual with a question about growth

  Incorrect

  Q&A answers a question with a visual, not a written sentence you control.

• DA smart narrative in Custom mode, with dynamic values

  Correct

  Custom mode lets you write the text and insert values tied to fields and measures.

The narrative visual has a Copilot mode, which generates summaries with AI, and a Custom mode, where you write and format the text yourself and insert dynamic values tied to fields and measures, so the narrative updates with the data.