> For the complete documentation index, see [llms.txt](https://ailyze.gitbook.io/ailyze-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ailyze.gitbook.io/ailyze-docs/analyze-data/discover-themes-and-patterns.md).
# Discover Themes & Patterns
This page shows you how to let Evidano surface the recurring [themes](/ailyze-docs/reference/glossary.md#theme) in your text, then shape them into a set you can defend with quotes and counts.
{% hint style="info" %}
**What you need:** a document or spreadsheet project you've analyzed — interviews, open-ended survey answers, reviews, or reports.
**What you will have at the end:** a reviewed set of themes and sub-themes, each backed by quotes and frequency counts.
{% endhint %}
Thematic analysis is the backbone of qualitative research. Evidano does the first pass; you keep the judgment.
## Three ways to get themes
When you reach the themes step (in document or spreadsheet analysis), you choose how the themes are created:
1. **Let Evidano generate them** — the AI proposes a [codebook](/ailyze-docs/reference/glossary.md#codebook) of themes and sub-themes from your data. Best for exploration and speed.
2. **Enter your own** — type or paste a codebook you already have. Best when you're applying a framework you bring to the data.
3. **Reuse a previous codebook** — choose **"Use Previous"** to apply the codebook from an earlier project, keeping studies consistent.
{% hint style="info" %}
**Tip:** You can steer the AI. In the **AI Theme Generation** dialog, add an instruction like *"I want 3–5 themes focused on barriers, motivations, and trust; keep sub-themes concrete."*
{% endhint %}
These first two options mirror two classic approaches: letting themes emerge from the data ([inductive](/ailyze-docs/reference/glossary.md#inductive-coding)) or applying a framework you already have (deductive). Most real projects are a hybrid — start with the AI, then edit toward your framework. For the reasoning behind each, see [Choosing the right analysis approach](/ailyze-docs/methods-and-best-practices/choosing-the-right-analysis-approach.md).
## Refine before you commit
Evidano shows the proposed themes so you can **edit them before running** the full analysis:
* Rename vague themes to your study's language.
* Merge near-duplicates; split themes that combine two ideas.
* Add a missing theme or sub-theme.
* Add **descriptions** so coding stays consistent (there's an AI helper for this).
You can also restructure themes *after* results using the codebook editor — see [Build a codebook](/ailyze-docs/analyze-data/build-a-codebook.md).
## Reading the patterns in your results
Once analysis completes, look beyond the theme list:
* [**Frequency**](/ailyze-docs/reference/glossary.md#frequency) (codebook table and frequency charts) — what's most common.
* [**Co-occurrence network**](/ailyze-docs/reference/glossary.md#co-occurrence-network) — which themes appear *together*; reveals relationships and tensions.
* [**Hierarchy graph**](/ailyze-docs/reference/glossary.md#hierarchy-graph) — how sub-themes organize under themes.
* **Segment comparisons** — whether a theme is concentrated in one [segment](/ailyze-docs/reference/glossary.md#segment) (see [Compare groups (segments)](/ailyze-docs/analyze-data/compare-groups-segments.md)).
* **Quotes** — the evidence; always read them, don't trust counts alone.
For a definition of each visualization, see [How to read your results](/ailyze-docs/analyze-data/how-to-read-your-results.md).
## Quality checklist
* [ ] Do the themes actually answer your research question?
* [ ] Are themes distinct (low overlap) and named clearly?
* [ ] Does each theme have supporting quotes you've verified?
* [ ] Did you check for **disconfirming** evidence? ([Chat with your data](/ailyze-docs/analyze-data/chat-with-your-data.md) is handy here: *"What contradicts theme X?"*)
* [ ] If comparing groups, are differences real or just sample-size artifacts?
## A realistic example
> **Scenario:** 25 interviews on remote-work experience. Let Evidano generate themes, then merge "isolation" and "loneliness," split "tools" into "collaboration tools" versus "scheduling tools," and add descriptions. Run it, then use the co-occurrence network to show that *"isolation"* and *"manager check-ins"* rarely appear together — a finding worth a paragraph.
## Next steps
* [Build a codebook](/ailyze-docs/analyze-data/build-a-codebook.md) — curate and validate your themes and codes.
* [Compare groups (segments)](/ailyze-docs/analyze-data/compare-groups-segments.md) — see which groups a theme belongs to.
* [How to read your results](/ailyze-docs/analyze-data/how-to-read-your-results.md) — decode every chart and section.
* [Qualitative research best practices](/ailyze-docs/methods-and-best-practices/qualitative-research-best-practices.md) — the methods behind good themes.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://ailyze.gitbook.io/ailyze-docs/analyze-data/discover-themes-and-patterns.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.