Upload and analyze CSV surveys

Marvin can read survey data that is stored in CSV files and add it to your project. You can analyze responses, make notes on them and add participant data to the research panel.

Multiple apps allow you to export survey data in CSV files. Follow the steps in this article to include CSV survey files in a project in Marvin.

Please navigate to the project to which you want to add a survey.

You’ll see guidelines on how to prepare your CSV for upload. Once the file is correctly formatted, drag and drop a file, or select "click to upload".

You can choose to import data from specific or all questions in the survey. Each question is imported as a separate column.

If you want to remove columns, uncheck the ones you want to leave out.

To begin, you need to identify the columns that contain participant data, such as names and email addresses. Marvin will auto detect the column using AI.

In case it hasn't detected it correctly, click "Select from all columns". Select any other column that you want Marvin to consider as responses for participant names

If you don't have the names of respondents, click "Survey doesn't contain respondent names" to proceed.

Similarly, Marvin will detect the column that contains emails.

To choose another column, click "Select from all columns".

If you don't have respondent emails, click "Survey doesn't contain respondent emails" to proceed.

You can also add these participants to your Research Panel in Marvin. We recommend you do this to maintain your participant database. You can also keep track of all the interactions each user has with the different teams and channels in your company.

Next, you’ll see an option to specify the date and time when the survey was created.

Marvin will use AI to automatically categorize each question in the survey as open-ended, NPS, single choice, card sort, and ranking.

You can check the AI mapping for each question. Marvin will highlight the ones where you need to confirm the categorization. If Marvin has correctly identified the type of question, click"Yes". To change the categorization, click "No" and choose the right type of question.

Complete the import once you've checked all the questions. You can also check and revise after the survey has been uploaded.

It may take a few minutes to complete the import. The task will continue running in the background even if you navigate away. You'll find the survey in the project files once the upload is complete.

Marvin presents all the survey data in three tabs – Summary, Individual Respondents and, Responses.

You’ll see all the data grouped by question in the Summary tab.

Use this tab to view all the responses of one participant in a single table. Easily navigate between respondents by clicking on their name

To view all responses in one, easy-to-read table, head to the Responses tab. You can add file tags and edit data directly from the table.

You have a few options to filter and analyze the survey data.

Click the drop down menu against "Filter responses by" to see filters for individual questions. You can choose to view data of specific responses to questions. You can apply multiple filters here. For example, you can focus on a data set of users who "Use your product twice a week" and are "Windows users".

Similarly, you can choose to see only the single-choice, multiple-choice, NPS or ranking responses when you click the drop down menu that says "Single choice".

Marvin can arrange all the metadata or quantitative data in a bar graph or pie chart (or notes in case of open-ended questions) when you click "Pie Chart". You'll see the open-ended responses or qualitative data as notes in case of open-ended questions. Use these notes for analysis and cross-project analysis.

You’ll also have an option to add a survey question as a field to the research panel.

To derive insights, type in any questions and prompts into the Ask AI field. Marvin will search across questions to find you relevant answers.

You can use this to identify trends, patterns and outliers. Use it to speed up your analysis, focus on specific questions or participants, and surface hard-to-identify insights.