What you will be able to do
- Turn a worksheet query result into a Snowsight chart and change its type, columns and styling
- Use chart bucketing and aggregation functions to show trends without rewriting the query
- Explore relationships and distributions with Snowsight chart types and the worksheet statistics pane
- Present an analysis to business users as a Snowsight dashboard built from query-driven tiles
Key concept
Query-driven tile — In Snowsight, every chart on a dashboard is a view of one SQL query's result set. When you change the chart, filter it or troubleshoot it, you are working on that query and the context it runs in.
1.From query result to chart
Every visualization in Snowsight starts with a SQL query. You open a worksheet, run it, and select Chart above the results table. Snowsight then builds a first chart for you from the columns the query returned. The documentation describes the purpose like this: charts turn query results into visuals that communicate logical relationships and support better decisions. For the exam, this is what 'present data for business-use analyses' means. A business reader gets a chart that answers a question, not a grid of rows to scan.
Snowsight supports five chart types: bar charts, line charts, scatterplots, heat grids and scorecards. Two facts limit what you can do with them. First, each query supports one chart type at a time. If a stakeholder wants the same result as both a line and a scorecard, you need two presentations, not one chart with two modes. Second, the auto-generated chart is only a draft. You refine it in the side panel, and the table below shows where each change is made.
| Control | What it changes |
|---|---|
| Chart type selector | Switches the visualization, for example from a line chart to Bar |
| Data section | Adds or removes columns, swaps in a different result column, and changes how a column is represented, such as its bucketing |
| Appearance section | Styling. The available settings depend on the chart type, for example a heat grid |
| Hovering over the chart | Shows details for each data point |
Checkpoint 1 of 5· Check yourself
An analyst wants a single worksheet query's revenue result to appear as both a line chart and a scorecard at the same time, in one chart. What does Snowsight support?
A query drives one chart type at a time. You can switch the type with the chart type selector, but one chart cannot show two types at once.
“Each query supports one type of chart at a time.”Source: docs.snowflake.com
Sources1
2.Trends: bucketing and aggregation
A trend is change across an ordered dimension, usually time. Daily data is often too noisy to show the direction a business is moving, and the same result can look very different at a coarser grain. Snowsight charts handle this with buckets, which group roughly continuous numbers, dates and timestamps. Because the bucketing happens in the chart and not in the SQL, you can try daily, weekly and monthly views of one result set until the pattern is clear.
| Column type | Buckets available in the chart |
|---|---|
| Date column | date, week, month, year |
| Numeric column | integer values |
When a bucket holds many rows, the chart has to reduce them to one plotted value with an aggregation function. Snowsight offers average, count, minimum, maximum, median, mode and sum. Which one you pick is part of the analysis. A monthly sum of revenue shows a trend in volume. A monthly count of orders shows a trend in activity. When a few extreme values fall into a bucket, the average moves toward them, so the bar no longer shows a typical row. Median is the function in this list that reports the middle of the bucket. Charts help you find both patterns and outliers, so check whether an outlier is driving what a bar shows before you present it.
Checkpoint 2 of 5· Check yourself
Which bucket options does a Snowsight chart offer for a date column?
Chart bucketing for date columns covers date, week, month and year. Quarter is not on that list, and integer buckets apply to numeric columns.
“Charts can bucket by date, week, month, and year for date columns. For numeric columns, charts can bucket by integer values.”Source: docs.snowflake.com
Checkpoint 3 of 5· Exam question
A marketing analyst has a Snowsight worksheet result with daily `ad_spend` and `revenue` columns for 400 days. A director asks whether higher spend goes with higher revenue. Which chart type shows this relationship most directly?
Correct answer: C — A scatterplot with `ad_spend` on the X-axis and `revenue` on the Y-axis, so each day is one point and the slope and spread are visible
- A. A heat grid is meant for categorical or bucketed axes and shaded aggregates; it hides the individual point pattern a scatterplot reveals.
- B. A scorecard summarizes one metric; it cannot show how two variables move together across 400 rows.
- C. A scatterplot plots one numeric variable against another per row, which exposes direction, strength and outliers of a relationship.
- D. A bar chart of one measure by date shows spend over time, not its relationship to revenue.
Sources1
3.Correlations and distributions
A correlation question asks whether two variables move together, for example discount size and order value. In Snowsight, the first step is in SQL: return both variables as columns of the same result. Then pick a chart type in the chart type selector and use the Data section to choose which result columns the chart plots. Snowsight's chart types include scatterplots and heat grids alongside bars, lines and scorecards.
Be aware of a limit in the documentation used for this lesson. The Snowsight charting page lists these chart types but gives no per-type rules for when to use each. It also says nothing about correlation coefficients. What it does say is that charts let you quickly find patterns and outliers. Treat any scatterplot or heat grid view as an exploration of that kind, not as a statistical test.
Before you chart a relationship, it is worth checking the shape of each variable. When you select columns, cells or ranges in a worksheet result, Snowsight automatically generates contextual statistics in the inspector pane. Every column shows how many rows are filled and empty. If many rows are empty, an apparent relationship may rest on only part of the data. Date, time and numeric columns get a histogram of how many rows fall into each range. You can click a bar or drag across the histogram to select a range. Categorical columns, meaning text columns where values repeat, get a frequency distribution. Together these show skew and outliers early, which tells you which aggregation function and bucket size the chart should use.
Checkpoint 4 of 5· Check yourself
You select the numeric column ORDER_VALUE in a worksheet result. Which view does the statistics pane show for it?
Histograms are shown for date, time and numeric columns. Frequency distributions are for categorical text columns, and key distributions are for JSON columns.
“Displayed for all date, time, and numeric columns.”Source: docs.snowflake.com
4.Presenting the analysis as a dashboard
A single chart answers one question. A dashboard puts several answers in front of a business audience at once. Each tile on a dashboard is a chart or table driven by its own query, and readers can hover over charts to see each data point. You can create an empty dashboard from Projects » Dashboards by selecting + Dashboard. You can also promote a finished worksheet into a new dashboard.
Promoting a worksheet has a side effect that catches people out. The worksheet is removed from the worksheet list and can then only be reached from the dashboard, where its query is stored and can still be edited. If you had shared that worksheet, those users lose access to it and its links stop working.
Once the dashboard exists, the context selector sets the role and warehouse used to run the tile queries, and Run reruns all of them. New tiles are added at the bottom, and you drag tiles to arrange them in the order you want to present them.
Checkpoint 5 of 5· Put it in order
Put the steps for creating a dashboard from an existing worksheet in order.
- 1.Enter a name for the dashboard and select Create Dashboard
- 2.Select New dashboard
- 3.Hover over the worksheet name and select Move to
- 4.Open the worksheet
You move the worksheet into a new dashboard and give the dashboard a name. The dashboard then opens with a tile built from that worksheet.
“Enter a name for the dashboard, and then select Create Dashboard. The dashboard opens, displaying a tile based on the worksheet you used.”Source: docs.snowflake.com
Sources3
Exam traps
Each one states something that sounds right. Open it to see what is actually true.
1.To show a daily result as a weekly or monthly trend, you have to rewrite the query's GROUP BY.Why is that wrong?
Time bucketing is a chart setting in the inspector panel. You can change the grain without editing the SQL.
Covered in Trends: bucketing and aggregation
2.Moving a shared worksheet into a dashboard keeps it available to the colleagues it was shared with.Why is that wrong?
The worksheet is removed when the dashboard is created. Its permissions are revoked and its links stop working.
Covered in Presenting the analysis as a dashboard
Sources
Every claim above is drawn from one of these pages, quoted as it was written on the date shown.
- 1.
“Charts transform your query results into visualizations that communicate logical relationships and lead to more informed decision making.”
↩︎ From query result to chart“Chart generation and data transformations in worksheets can result in compute usage.”
↩︎ From query result to chart“Style your chart in the Appearance section. The available settings depend on the type of chart.”
↩︎ From query result to chart“Charts use aggregation functions to determine a single value from multiple data points in a bucket.”
↩︎ Trends: bucketing and aggregation“Charts let you quickly identify and understand patterns and outliers in data.”
↩︎ Correlations and distributions“Without modifying your query, you can easily select a different bucket of time (e.g. weekly or monthly data) in the inspector panel”
↩︎ Exam trap 1“Each query supports one type of chart at a time.”
↩︎ Checkpoint“Without modifying your query, you can easily select a different bucket of time (e.g. weekly or monthly data) in the inspector panel”
↩︎ Prediction“Charts can bucket by date, week, month, and year for date columns. For numeric columns, charts can bucket by integer values.”
↩︎ Checkpoint - 2.
“Contextual statistics are automatically generated for all column types. The statistics are intended to help you make sense of your data at a glance.”
↩︎ Correlations and distributions“Categorical columns are text columns where the same values are used more than once.”
↩︎ Correlations and distributions“Displayed for all date, time, and numeric columns.”
↩︎ Checkpoint - 3.
“the worksheet is removed from the list of worksheets and can only be accessed from the dashboard.”
↩︎ Presenting the analysis as a dashboard“Use the context selector to specify the role and warehouse to use for running the queries in the dashboard.”
↩︎ Presenting the analysis as a dashboard“Tiles visualize data on your dashboards as charts and tables. Hover over charts to view details about each data point.”
↩︎ Presenting the analysis as a dashboard“Dashboards are flexible collections of charts arranged as tiles. The charts are generated by query results and can be customized.”
↩︎ Key concept“If the worksheet is shared with other users, those users lose access to the worksheet when you create a dashboard”
↩︎ Exam trap 2“Enter a name for the dashboard, and then select Create Dashboard. The dashboard opens, displaying a tile based on the worksheet you used.”
↩︎ Checkpoint