# Dash Lab: research dashboards built for your question

Dash Lab is a [Cassi.ai](https://www.cassiai.com) service delivered as a product: an interactive research analysis platform built to measure for each study. Generic charting tools deliver the charts on their menu, the same for every business. Dash Lab designs the dashboard around the client's question, with the charts and statistical analyses that question calls for: importance and performance matrices, perceptual maps, cluster profiles, driver analysis with confidence intervals, significance tests inside the tables, wave tracking, heat maps, funnels and one-page executive infographics. Filters recompute every chart and every percentage shows its base.

## Every study ends up in the same menu of charts.

Tracking, segmentation and ad hoc results reach the client through generic charting tools. They show what is in the catalogue well. The analysis the research calls for is rarely there.

- Topic: Delivering research results. A brand tracker, a segmentation study, a driver analysis of satisfaction. The numbers are ready. What is missing is a place where the people who decide can explore them.
- How it is done today: A generic charting tool, plus a deck. Results go into a general-purpose dashboard tool that offers bars, lines and pies from a catalogue. What the tool cannot draw, a perceptual map, a significance test, a cluster profile, stays in the static deck or in a spreadsheet.
- The pain: The analysis that matters stays outside the panel. The analyst bends the question to what the tool can show. The importance and performance matrix becomes two tables. Significance is checked by hand. And the client receives a panel that looks like every other supplier's.
- Where Cassi.ai comes in: A platform built for the study. Each dashboard is designed around the client's question, with the charts and statistical analyses that question requires, from the matrix to the executive infographic, with their brand and their cuts.

## The analysis the study calls for, inside the dashboard.

- Matrices ("Which attribute deserves investment first?"): An importance and performance matrix with quadrants, and a perceptual map that places brands and attributes on the same plane. Each chart is drawn for the reading, with the base and the cut beside it.
- Segmentation ("Who are our segments and how do they differ?"): Each cluster's profile on a radar chart, the market composition on a stacked bar, and a segment card with the attributes that set it apart. Global filters cut through all of it.
- Driver analysis ("What weighs most on satisfaction or intent?"): Regression or relative importance per attribute, with the coefficient and its confidence interval drawn as an error bar. The reader sees what weighs and how much uncertainty sits in each bar.
- Built-in significance ("Is that difference real, or is it noise?"): T-test, chi-square or proportion test, defined with the research team and built into the table. The p-value sits in the cell, with markers for increase and decrease, and a low base is flagged and kept visible.
- Wave tracking ("What changed since the last wave?"): Reasons and indicators wave by wave, with the base of each period and the difference from the previous one tested. Global filters recompute every chart: 7 in the live example.
- Executive infographic ("How do I take this to the board?"): A one-page infographic assembled from the same data: big numbers, a map by region, a waterfall of price drivers and rows of icons. It leaves the dashboard ready for the meeting.

## Six steps, from the business question to the next wave.

1. Start from the question. We map the decisions the study has to support and who will open the dashboard. The layout follows the question.
2. Read the data the study already produces. The dashboard is fed from the data your collection already generates. Variables, weights and bases are checked against your tables. The numbers match your tabulation before design begins.
3. Design layout, charts and analyses. Screens, filters, charts and statistical analyses are chosen for this study, and drawn from scratch when the method asks for it, with the client's brand. You approve the design before it is built.
4. Build the statistics in. Significance tests, confidence intervals, base sizes and low-base flags go into the charts and tables themselves. No difference is marked without the test.
5. Validate with the research team. The team that ran the study goes through every screen and the filter combinations that matter. Nothing reaches the client unchecked.
6. Publish and update. Each client gets access to its own platform, which updates with each new wave or data load. One client's data is never visible to another.

## Numbers from the live example, stated with their scope.

- 7: global filters in the live example
- 12: weeks of tracking in the live example
- 6: journey stages in the live example
- <25: base size that raises the low-base flag in the live example

## What it never does

- Never shows a percentage without its base.
- Never hides a low base. It flags it and keeps the number visible.
- Never marks a difference as significant without the test.
- Never ships a generic template with the logo swapped.
- Never exposes one client's data to another.

## Who it is for

Research firms that deliver tracking, Insights teams, Brand and CX teams, Consultancies. For studies whose results need to be explored by the people who decide, wave after wave, with the analysis the research calls for and without a new request for every cut.

## Questions and answers

### What is Dash Lab?

Dash Lab is a Cassi.ai service delivered as a product: an interactive research analysis platform built to measure for each study. The dashboard is designed around the client's question, with the charts and statistical analyses that question calls for, its own layout, filters and tests, and the base next to every number.

### How is it different from Power BI, Tableau or Looker?

Generic charting tools, such as Power BI, Tableau and Looker, deliver what is on the menu: bars, lines, pies and tables, the same for any business. They serve operational reporting well. A Dash Lab platform is built for a research study: importance and performance matrix, perceptual map, cluster profiles, driver analysis with confidence intervals and significance tests are designed for the question, with the base beside each number. The same applies to the panel bundled with the collection tool.

### Where does the data come from?

From the data the study already produces, whether that is what your collection tool generates or your own database. Before design begins, variables, weights and bases are checked against your tabulation, so the dashboard shows the same numbers your team already trusts.

### How does significance testing work in the dashboard?

The test is defined with the research team for each study and built into the tables: t-test for means, chi-square or proportion test for percentages. The p-value sits in the cell, with markers for significant increase or decrease. In the live example, each weekly difference is tested against the previous week. No difference is highlighted without the test.

### What happens when a filter leaves too few interviews?

The dashboard keeps the number visible and flags it as a low base. In the live example the flag appears under 25 interviews. Every percentage is shown with its base, so whoever reads the chart knows how many people are behind it.

### Which analyses and charts does Dash Lab build?

Global filters and indicator cards, funnels, wave tracking, heat maps, treemaps, importance and performance matrices, perceptual maps, radar charts and stacked composition bars for segmentation, regression coefficients with error bars for driver analysis, panels of t-tests and chi-square tests, waterfall charts, maps by region and one-page executive infographics. The list grows with the study: if the question asks for a chart that is not in any catalogue, it is drawn.

### Can the dashboard follow our brand and our client's brand?

Yes. Layout, colors and typography are designed per project. A research firm can deliver the platform under its own brand or under the end client's brand. Dash Lab never ships a generic template with the logo swapped, and one client's data is never visible to another.

### What can I see in the live example?

An automotive market research showcase with fictional data: 7 global filters, indicator cards by segment, a purchase abandonment analysis across 6 journey stages, 12 weeks of reasons with base size, weekly difference and significance, a treemap of reasons and a heat map of stage by reason. The matrices, the segmentation, the statistical tests panel and the infographic appear in the animated preview on this page.

## Bring the question your current dashboard cannot answer.

Send the question the study has to answer and a sample of the data it already produces. We come back with a layout, the charts and the analyses it would support. https://dashlab.cassiai.com/#acesso