Start by importing your data into Yabble
Gen is the ultimate AI Research Agent that allows you to have a conversation with your own data – all within Yabble’s secure, walled garden data environment. Gen uses both a natural language model and SQL queying which means you can ask your structured and unstructured data all types of questions and Gen will be able to answer instantly and conversationally. Once your dataset is uploaded into the Yabble platform you can prompt Gen to query your dataset and provide high level insights or dive deeper to explore your data holistically – all with data security peace of mind.
How to import data for Gen chat without theming and coding analysis
1. Prepare your data for import
Are you ready to import your short form data to chat with Gen? Follow our easy checklist to ensure your data file will process successfully.
Ensure your file follows these data limits – your import may fail if these are not met:
There is a maximum limit of 9000 characters for each text response being analyzed.
There is a maximum of 100 custom variables per file.
A date column is mandatory and there can only be one date column per file.
Use Row 1 for headers.
For custom variables – we recommend making them short and easy to understand eg. Date, Location, Gender. These labels are used for navigation and filtering of data. Avoid labels not all users of the data may know e.g Q1/A1 etc.
For text data – ensure column titles are descriptive to guide the AI on how to count your data. For example, “How would you describe a restaurant that was good value for money?” as opposed to just “Good value restaurants”. This helps to give the AI context and better theme your responses.
Top Tip – do not just use the survey question as the label. As a survey question does not always provide enough context for the AI.
Clean your data file. Clear your file of junk and gibberish responses. This will ensure you’re only paying for insights from high-quality data. Our quality control processes also automatically look for these types of irrelevant responses and remove any identified. However, the best quality control is your own.
eg. xyz, qwerty, nah, idk
eg. bot generated responses not relevant to your topic.
Add a date column and use the digestible format (dd-mm-yyyy hh:mm:ss or dd-mm-yyyy). Every upload requires a date column. Use your respondents’ completion date or the survey date, or select the day of your import if the date isn't relevant to you. Inputting an accurate date now will allow use of the trending feature for longitudinal data.
Add custom variables; they’re free! You can upload up to 100 custom variables These help you filter your data for deeper analysis once it’s been processed.
What are custom variables? Custom variables are additional pieces of information to assist with your analysis and are particularly useful for filtering your Counted data. Common examples include demographics, attitudes and ratings.
Save your file as a CSV. Make sure you're uploading CSV files only as this is the only format accepted.
Once you’ve followed all these steps and your file is ready to upload, you can proceed with your import.
2. Import your data
Accessing the import feature
If you are on the Yabble home page, navigate to the Projects page, then select ‘New project’ > 'Theme Counting’.
Data Import Limits
For each data file uploaded there are some limits which if exceeded will cause the file upload to fail.
There is a maximum limit of 9000 characters for each text response being analyzed.
There is a maximum of 100 custom variables per file
A date column is mandatory and there can only be one date column per file.
Project Labelling
We recommend descriptive names of not more than 4-5 words to make them easy to select and navigate to when inside the platform.
Naming Recommendations for Tracking Projects
Project Name: The first time you upload data for a tracking project we recommend the project name clearly indicates the project is a tracker. This will enable it to be easily found for future uploads of data by you and other users. E.g. Online Shopper Satisfaction Tracker.
Import Name: We recommend labelling using the relevant time period. E.g. Online Satisfaction Tracker Q1 2024. Then for future uploads the import names can be sequential e.g. Online Satisfaction Tracker Q2 2024 etc.
Naming Recommendations for One Off Projects
Project Name: This will be how you search and select the project so use descriptions that make the most sense to you.
Import Name: This is less important for one off projects vs trackers so repeating the project name is fine
Uploading your Data
This step helps to contextualize your data, allowing us to ensure we apply the models that will generate the most accurate output from your dataset.
Select the appropriate Industry from the dropdown box and give your import a name to help you easily keep track of and combine files.
Drag and drop or browse and select; Please note that only CSV files are supported
Tick the box to confirm that your file doesn’t contain any legal, medical, or political data and that you have full permission to use it. (For more details, view our Data Policy.) When you're ready to map your columns, click next.
Mapping your data columns when importing
Here, you’ll review the columns in your dataset to confirm how you’d like to proceed with your upload.
Important note: you must have a response date column to successfully upload. Only one response date column is allowed per file. See the ‘Prepare your data for import’ tutorial for tips on how to prepare your file for upload.
If you do not want to theme count your data, make sure each column is unticked/deselected. You can also remap column types between the following: Open text, Custom variable, NPS and Response date.
When selecting your open text questions, giving them context is helpful in getting the best output. Typically, in a survey we might follow-up a rating question with ‘What is the reason for your rating?’. To get more specificinsights from Gen, you might consider changing this in the data mapping stage to something like ‘Why are you likely or unlikely to recommend <brand> to your friends or family?’, or ‘Why are you likely or unlikely to continue using <brand> as your energy supplier?’
To rename a column, select the text box under column name and type in your preferred name.
Select the columns you’d like to run Hey Yabble Count on.
You can select all columns by ticking the top left corner box. Deselect any columns you don’t wish to count. We only apply counting to valid data points where no errors or null values have been detected.
Note: If you do not want to theme count this data, and only want to chat with Gen, make sure all columns are deselected.
To complete your import, review your details and available credits. Once you’re ready, hit Complete to launch your data upload.
How to import data for Gen chat with theming and coding analysis
Importing your short-form data to chat with Gen follows the same process as importing data for Count – if you want to chat with your theme counted data, see our Count module for tips on importing short-form data to the Yabble platform.
How to import long-form data for Gen chat with AI summarization
Importing your long-form data to chat with Gen follows the same process as importing data for Summarize – if you want to chat with your long-form data, see our Summarize module for tips on important long-form data to the Yabble platform.
Once you have imported your data and it has been processed in the system, you can start chatting with Gen – see our 'Start chatting with Gen' module to learn how.