> ## Documentation Index > Fetch the complete documentation index at: https://conveo.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Dashboard & metrics > Build Storyline charts and interpret their populations, percentages, and comparisons over time. Use a Storyline **Dashboard** to follow measures and compare groups over time. A chart supplies the counts or trends behind a finding; read it alongside the [insight's evidence](/docs/storylines/insights). ## Metrics A chart uses an available analysis measure or category, called a **facet** in the chart builder. Examples include a coded topic or a numeric answer. Select the values you want to track, then choose a visualization. Open **Dashboard**, choose **Edit**, then **Add card manually** to open **New chart**. Configure the measure, visualization, optional grouping or split, and optional audience filter. Choose **Add to dashboard**, then **Done** when you finish arranging the board. Changing the dashboard requires edit access to the underlying study. ### Choose the population and measure | Setting | Behavior | | ------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Completed interviews** | Counts completed interviews, using completion dates for calendar charts and reporting waves for wave charts. | | **All participants** | Includes people who entered the study, including screened-out, abandoned, canceled, and in-progress interviews. Uses entry dates and entry waves. Errored interviews and CSV invitations with no participation are excluded. | | **Share of respondents** | Counts respondents with the selected values; use the percentage setting when you want shares. | | **Average value** / **Median value** | Summarizes numeric values for a numeric measure. These choices are not available for every facet. | Both populations exclude test, hidden, and deleted interviews. **All participants** also excludes simulated interviews. Talk to your Data uses completed interviews, so it does not provide the same all-participant recruitment view. Switching the population can change both the count and the period an interview appears in. Use the same population when comparing charts. ### Percentages and missing data For a measure tied to a question, percentage calculations use interviews where the question was not skipped and at least one value was extracted. An interview with no extracted reading is not automatically a zero. For a measure that is not tied to a question, the percentage base includes the selected population; no detected value can count as not mentioning that category. Different measures can therefore have different bases. Check data coverage before treating a change in percentage as a change in opinion. A **split** compares groups separately. An audience **filter** narrows who is included. These settings answer different questions: comparing regions is different from restricting the chart to one region. ## Dashboards A Storyline can contain multiple dashboards. Use a separate board for a focused audience or research question. In edit mode, arrange cards and use **Add section** to group them. Select **Group by** to compare reporting waves or calendar periods. Wave grouping displays the reporting waves; the calendar period selector is hidden in this mode. The default follows each chart's saved grouping. Some pinned charts are snapshots, such as wave-to-wave transition charts, and do not recalculate into a different time grouping. Open a chart's title for its detail view. Before sharing a result, check the selected measure, population, period, grouping, and filter together. ### Manage boards and cards Use **Add dashboard** beside the board tabs to create another board. Double-click the active tab to rename it. The delete control is available when more than one dashboard exists; the final board cannot be deleted. Deleting a dashboard permanently removes its charts and sections. Deleting only a section divider leaves the charts intact. Review the confirmation before removing a board. In edit mode, hover over or focus a card to reveal **Configure card**, **Duplicate card**, and **Delete card**. A pinned scatter snapshot cannot be reconfigured in the manual builder, but can be duplicated or deleted. Its original cross-measure comparison remains fixed. For eligible cards, the **%** and **#** controls switch the displayed percentage or interview count. This is a viewing choice; configure the card to change its saved default. ### Visualization options The manual builder offers **Amount**, **Line chart**, **Bump / rank**, **Stacked area (100%)**, **Stacked bar**, **Grouped bar**, **Horizontal bar**, **Pie chart**, **Heatmap**, **Ordinal scale**, **Crosstab**, **Waterfall**, and **Scatter**. Choose a form that matches the question: a trend over time, a category comparison, or a relationship between values. A manually built **Scatter** requires exactly two values from the chosen facet. To compare values from two different facets, use a pinned correlation signal instead. Not every chart supports the same settings. **Running total** is available for line charts counting respondents. It accumulates counts over the displayed periods and turns off percentage display. It does not apply to average or median measures. ### Group, trim, and label values Use **Group values** to combine categories in the chart, or **Split by facet** to compare separate groups. These are chart settings; they do not recode the source answers. **Trim long tail** limits the categories displayed. Choose **All**, the top **5**, **10**, **20**, or **50**, and optionally a **Min %** threshold. Trimming hides smaller categories from that chart; it does not establish that those views were absent. Check an untrimmed view before reporting the full range of responses. Use the title and available axis, category, and split-label fields to make the chart understandable to colleagues. Keep labels faithful to the underlying measure. A friendlier label does not change what was counted. ## Why this matters A number can move because opinions changed, different people participated, or fewer interviews produced a usable reading. Compare equivalent populations and question definitions, and inspect the supporting interviews before deciding what caused a change. Small groups and sparse periods can give unstable percentages. A dashboard does not guarantee statistical representativeness or a causal explanation. Live charts can also change as their underlying data changes; a saved report or chart snapshot is not necessarily the latest calculation. ## The Overview The former **Overview** page has been retired. Its address redirects to the program's current entry point. Use **Dashboard** for charts and **Insights** for findings. Accounts with **Highlights** and **Explore** also have those views for investigating changes; they are not available to every account. ## Build and revise a dashboard with the assistant Use the dashboard assistant to request charts or changes in plain language. For example: “Add a weekly satisfaction chart and place it under Experience.” The assistant can propose chart additions, configuration changes, duplication, moves, resizing, and section dividers. Review proposed changes and accept or reject them individually or together. A preview is not a saved change until accepted. Check the proposed measure, population, filters, and period before accepting: an AI-generated chart still needs the same interpretation checks as a manually configured chart. The assistant can also propose dashboard tabs and move a chart between dashboards within the same Storyline. Moving a chart removes it from the source board and places it on the destination; use duplication when you want a separate copy. The assistant can delete only an empty dashboard, and cannot delete the final dashboard. Manual dashboard deletion has its own confirmation and can remove a board with content.