Visualization types
Each example below shows a visualization built in DEAP Explorations next to the configuration used to create it. The Variables tab lists the variables and the grouping, coloring, and stratification selected; the Settings/Notes tab lists the chart settings and any adjustments made in the Text or Colors tabs. To learn more about each option see Customization.
Bar, dot, & lollipop charts
Bar charts, dot plots, and lollipop charts show the count or percentage of responses at each level of a variable. They are best for comparing frequencies across categories, such as the number of participants per site, the distribution of survey responses, or how response patterns differ between sessions or participant groups.
- Variables: Select categorical or numeric variables. When using numeric variables, we suggest only using those with a
type_levelof nominal, ordinal, or interval. - Single or multi-variable: Plot one variable on its own or several together (see Multi-variable plots).
- Adjustments: Grouping, coloring, and stratification by row or by column.
- Grouped or stacked: Display bars side by side (grouped) or on top of one another (stacked).
- Flip axes: Swap the X and Y axes, for example to display bars horizontally.
- Synchronize axes: Keep the same axis scale across stratified plots so they can be compared directly.
| Variable(s) |
ab_g_dyn__design_site
|
| Grouping | Categorical levels |
| Coloring |
By variable: ab_g_dyn__visit_type
|
| Stratification (column) |
By variable: ab_g_stc__cohort_sex
|
| Flip axes | True |
| Synchronize axes | True |
| Show as | Count |
| Notes | Text settings are modified to allow large Y-axis labels. Labels are right adjusted and then have ‘200’ padding. |
Scatter chart
Scatter charts plot two continuous variables against each other, with one point per participant-session. They are best for examining the relationship between two measures, such as spotting trends, clusters, or outliers, or seeing whether that relationship differs between participant groups.
- Adjustments: Coloring and stratification by row or by column.
- Load all data points: By default 20,000 data points are loaded. Check Load all data points to plot every participant-session; for large datasets this may significantly increase load time.
- Flip axes: Swap the X and Y axes.
- Synchronize axes: Keep the same axis scale across stratified plots so they can be compared directly.
| Variable(s) |
X axis: ph_y_anthr__weight__r01_001 Y axis: ph_y_bld__rslt__chol_qnt
|
| Coloring |
By variable: ab_g_stc__cohort_sex
|
| Stratification | N/A |
| Flip axes | False |
| Synchronize axes | N/A |
| Notes | All data are displayed in the plot (set below the Y Axis) but the plot is filtered to only show two sessions of data. The size of the key has also been adjusted in text settings. |
Histogram
Histograms show the distribution of a continuous variable by sorting its values into bins and counting how many fall into each. They are best for checking the shape, spread, and skew of a measure, spotting outliers or implausible values, and comparing distributions between sessions or participant groups.
- Single or multi-variable: Plot one variable on its own or several together (see Multi-variable plots).
- Coloring: Color by a variable, and filter which of its values are included in the coloring.
- Stratification: Stratify by row or by column.
- Chart type: Display a histogram, a density curve, or both.
- Show as: Display values as a count or a percent.
- Flip axes: Swap the X and Y axes.
- Synchronize axes: Keep the same axis scale across stratified plots so they can be compared directly.
| Variable(s) |
ph_y_anthr__height__r01_001
|
| Coloring |
By variable: ab_g_stc__cohort_sex
|
| Stratification (row) |
session_id
|
| Chart type | Histogram (No density) |
| Flip axes | False |
| Synchronize axes | False |
| Show as | Count |
| Notes | Chart is recolored manually (Colors tab) and key has been moved from default position and size increased (Text tab) |
Summary table
Summary tables report descriptive statistics for one or more variables in a table. For continuous variables these include statistics such as the mean, standard deviation, minimum, median, maximum, and the number of complete and missing values; for categorical variables they show the count of each response. They are best for a quick numeric overview of several variables at once, including how much data is missing.
- Stratification: Stratify by one or more variables.
- Flip axes: Swap the rows and columns of the table.
- Displayed as columns: Choose whether the table columns show the focus variables or the stratification levels.
Count table
Count tables list every combination of levels across the selected variables, with the number of participant-sessions in each. They are best for checking sample sizes, such as how many participants were seen at each site, in each session, or by each visit type.
- Rearrange variables: Grab a variable by its handle () and drag it up or down to move it in the list. The order of the list sets the order of the table columns.
- Show level labels: Choose whether categorical levels display as their response labels (True) or as their coded values (False).
- Transform variables: Use the Transform dropdown to transform each variable before it is counted:
- None: Count the variable’s values as they are.
- Is N/A: Count whether each value is missing.
- Not N/A: Count whether each value is not missing.
- Custom: Count whether each value meets a condition you set. Numeric variables use More than, Less than, Equal to, or Not equal to a value; categorical variables use Is in or Is not in a set of response levels.
Crosstabulation table
Crosstabulation (crosstab) tables cross the levels of one set of variables against another, showing the count and percentage of participant-sessions in each combination along with row and column totals. They are best for comparing how categorical variables relate to one another, such as the balance of sex or visit type across sites.
- Variables: Add up to two variables per axis (X1 and X2; Y1 and Y2). See Stratification for how crosstab dimensions differ from other plot types.
- Focus variable: Optionally add a focus variable and summarize it in each cell by its median, mean, minimum, maximum, or standard deviation.
- Display as: Show each cell as a count, a percentage, or both.
- Calculate percentage by: Calculate percentages by row or by column.
Correlation table
Correlation tables show the pairwise correlations between the selected variables as a matrix, with optional heatmap coloring and p-values. They are best for screening relationships across many variables at once, such as checking how closely parent and youth reports of the same measure agree.
- Correlation mode: Correlate between variables, or within variables (each variable with itself across sessions).
- Stratification: Stratify by one or more variables.
- Correlation type: Spearman (Pearson coming soon).
- Heatmap colors: Color each cell by the strength of its correlation, or leave cells uncolored.
- Include p-values: Show (True) or hide (False) the p-value in each cell.
- Cell visibility: Show only cells whose correlation falls within a range between -1 and 1, with the option to synchronize the upper and lower bounds.
- P-value filtering: Set a p-value threshold between 0 and 1; cells with p-values below that value are filtered out.













