> For the complete documentation index, see [llms.txt](https://evidano.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://evidano.gitbook.io/ailyze-docs/analyze-data/analyze-a-spreadsheet-with-evidano-stats.md).

# Analyze a Spreadsheet With Evidano Stats

Analyze a survey or monitoring spreadsheet by chatting with the AI in Evidano Stats (beta): clean the data, get tables, charts and coded themes, then export a report.

This page is for anyone who wants to analyze a spreadsheet by talking it through with the AI instead of filling in a form. Evidano Stats (beta) reads your data files and any questionnaire or terms of reference, cleans the data with you, produces tables, charts and coded themes, and drafts a report you can export. It sits beside the [spreadsheet analysis](/ailyze-docs/reference/settings/spreadsheet-analysis.md) setup, which still works as before.

{% embed url="<https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-ab4e0c67296acbe3fd9b833e9320191af7325434%2Fanalyze-evidano-stats-01-tour.mp4?alt=media>" %}
Video of steps 1 to 8 with an invented clinic survey (no sound)
{% endembed %}

{% file src="/files/wy8PSttt5Lndd28GkrX5" %}
Download this video
{% endfile %}

## Before you start

* An Evidano account. The Free plan shows the first results and codes one open-ended question; the written report and exports need Pro. See [Plans, Limits & Credits](/ailyze-docs/reference/plans-limits-and-credits.md#key-per-project-limits).
* Your data as Excel, CSV, TSV or JSON. No data yet? Download the invented sample below; the screenshots and video use it.
* Optional: the questionnaire, codebook, terms of reference or proposal, as PDF, Word or text. The AI uses them to name variables and organise the report around your questions.

{% file src="/files/b5hAU2MGcKRO07QY0E7p" %}
Sample survey: 60 patients of a mobile clinic in two districts (invented)
{% endfile %}

## Step 1: Open Evidano Stats

1. On your account page, find the **Spreadsheet analysis** card and click **New spreadsheet project**.
2. At the top right of **Upload/select files for analysis**, click **Try the new spreadsheet analysis (beta)**.

You should see **Let's start with your files**. **Use the old version** goes back to the spreadsheet setup, and **Your projects** lists your earlier Stats projects.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-fd021251fa26898e6cd1c3884143b2c1fc3321fe%2Fanalyze-evidano-stats-01-home.png?alt=media" alt="Evidano Stats start page with Clean, Analyze and Report cards and a drop zone with Choose files"><figcaption><p>Add every file at once: data, follow-up rounds and the questionnaire.</p></figcaption></figure>

## Step 2: Add your files

Drop your files on the page, or click **Choose files**.

You should see **Your files**, with the size of each file and the kinds of questions the AI found in it. **Add more files** adds another.

## Step 3: Choose how hands-on to be, and which columns

1. Under **How hands-on do you want to be?**, pick one:
   * **Autopilot**: the AI does everything and reports the decisions it made.
   * **Co-pilot**: the AI proposes and you approve the key decisions.
   * **Step-by-step**: the AI explains the options and you decide each step.
2. Under **Which columns should I analyze?**, untick the columns to leave out, such as respondent IDs. **Suggested** restores the AI's choice.
3. Click **Start analysis**.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-242330306e48cd2ad7b1e1b97f1ca52e359e765b%2Fanalyze-evidano-stats-02-your-files.png?alt=media" alt="Your files card with the hands-on choice, the column list with sample values, and Start analysis"><figcaption><p>The line under the file name says what the AI found: rating items, choices, open text, numbers, IDs.</p></figcaption></figure>

## Step 4: Answer the AI's questions

The AI checks answer codes, missing values and free text, then takes a first look at the data. When it needs a decision, it asks on a card in the chat, for example whether to keep to descriptive statistics or add significance tests (**Inferential**). The AI words each question itself. Click an answer, or type your own.

You should see the cleaning steps listed as the AI makes them. Every change can be undone from the workspace's **Cleaning log**.

You can change the hands-on level and the depth later with the two buttons at the top right of the chat.

## Step 5: Read the results

Each result arrives as a table or chart with one sentence on what it shows. Under each one:

* **Explain**: the AI explains it in plain language.
* **Add to report**: keeps it for the written report.
* **Open in workspace**: opens it beside the chat.
* **Copy table**, **Image** and **Data** copy or save it.

Ask for anything else in the box at the bottom, for example "Compare satisfaction by district" or "Chart the improvement themes by age group". Type `/` for a list of commands.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-ee9629c62e6213ecdb039efd74090a2ff7dbb9f4%2Fanalyze-evidano-stats-03-first-look.png?alt=media" alt="A frequency table and bar chart of respondents by district, with Explain, Add to report and Copy table below"><figcaption><p>The note under each chart says how to read it.</p></figcaption></figure>

## Step 6: Code the open-ended answers

For each open-ended question, the AI suggests a codebook on a card: themes, the codes under them, and a description of what each code covers.

1. Rename, move or delete codes, or click **Add code** or **Add theme**. The AI codes each answer by the descriptions, so say what counts.
2. Click **Code all answers**. **Reject** drops the suggestion.

You should see the themes by group and a table of quotes for each theme. Coding is free up to a word allowance per project; above it, the cost is shown and you confirm before anything is charged (see [Plans, Limits & Credits](/ailyze-docs/reference/plans-limits-and-credits.md#key-per-project-limits)).

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-f6dbea677f9f3a15ee3e303c9cea294b5f87195d%2Fanalyze-evidano-stats-04-codebook.png?alt=media" alt="Codebook card for What worked well with Levels, a theme, a sub-theme, and a description for each code"><figcaption><p>Levels switches between themes with codes and themes with sub-themes and codes.</p></figcaption></figure>

## Step 7: Draft the report

Ask for it in your own words, such as "Draft the report", or click the suggestion under the AI's last reply.

You should see the sections listed as they are written, each with **Open report**.

## Step 8: Edit and export the report

1. Click **Workspace** at the top right, then the **Report** tab.
2. Edit any section, or use **Draft with AI** to redraft one. **Style** sets the kind of report, for example **Evaluation report**.
3. Export with **Word**, **HTML** or **Markdown**.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-5117b98b4c5bbfe5984bcfc20f964b6755539869%2Fanalyze-evidano-stats-05-report.png?alt=media" alt="Workspace Report tab with Style, Draft with AI, the Word, HTML and Markdown buttons and a drafted Key findings section"><figcaption><p>The other tabs hold the data, every output, the analysis code and the codebook.</p></figcaption></figure>

Your project, its files and the conversation are saved to your account. Reopen it from **Your projects**; **New chat** starts another conversation on the same files.

<details>

<summary>Optional: the workspace in detail</summary>

* **Data View** and **Variable View**: the cleaned data, and each variable's label, type and answer codes. **Cleaning log** lists every change, each one undoable.
* **Output**: every table and chart; **Code**: the analysis script; **Plan**: the analysis plan and what is done.
* **Codebook** and **Sources**: the coded themes and the files you added.
* The menu bar (**File**, **Data**, **Transform**, **Analyze**, **Graphs**, **Qualitative**, **Report**) runs the same steps without the chat.

</details>

<details>

<summary>Optional: if something goes wrong</summary>

* **Start analysis stays grey:** the line beside it says what is missing, usually a file that is still being read.
* **The AI stops with "I paused: I kept repeating the same report section.":** click **Continue**, or open the **Report** tab; the sections already written are kept.
* **A card above is waiting for you:** scroll up and answer the card; the AI waits for it.

</details>

## Next steps

* Plan the analysis — [Survey Research Best Practices](/ailyze-docs/methods-and-best-practices/survey-research-best-practices.md)
* Code your own way — [Build a Codebook](/ailyze-docs/analyze-data/build-a-codebook.md) explains themes and codes
* Share what you found — [Export & Share Findings](/ailyze-docs/create-deliverables/export-and-share-findings.md)