CSV is the only file type that can be uploaded to the Keatext text analytics platform.
Your first step to importing data begins with a well-configured CSV file. To learn more about CSV files and preparing your data for upload, click here.
All of your uploaded data sources are accessible on the Sources page of the platform.
Adding a CSV file to your sources
Once your CSV file is properly prepared, you're ready to upload it for analysis. On the Sources page, add your new source from the files on your stored on your computer.
The app automatically detects the delimiter, the fields and the data types used. While typically accurate, it’s a good idea to double-check the data types when configuring the data fields.
Configuring your data fields
Once your CSV file is uploaded, configure the data fields so Keatext can interpret and analyze the data correctly.
Under “Configure fields” you'll see the system has automatically selected “comma” as the delimiter used in your file. If the delimiter hasn’t been correctly auto-detected, choose the right delimiter amongst the options in the pull-down menu.
You have the option of giving your dataset a source name: the original spreadsheet file name automatically appears in this field, but can be changed for clarity.
For future reference, you can add a brief description of your source. This can be a useful piece of information when working with multiple people in the same account.
If applicable to your dataset, choose a date field (we say "primary date" as you may have more than one - choose the one of most importance). This is how the Keatext system detects temporal trends.
You also have the option of choosing a commenter ID field (a unique customer/employee identifier such as an email or internal customer/employee ID linking an individual across channels). It will allow you to estimate the number of customers affected by an issue.
These settings cannot be modified once your source has been uploaded.
Mapping your data fields
First, you’ll need to map the data fields in accordance with the configuration of your original spreadsheet. The purpose of this process is to give Keatext accurate information about the original spreadsheet and to identify which spreadsheet fields should be analyzed.
Each field will automatically be named in correlation with the original spreadsheet's fields. Next to each field name, identify that field's data type (as in the original CSV) via a pull-down menu. The options are: text (string of characters, such as long form textual responses), number (a price or any other numeric value) and datetime (click here for all supported date formats).
Selecting fields for analysis
Perhaps the most important aspect of uploading data for text mining is identifying long form textual comment or feedback fields and check-marking them. In your original dataset, you can see which fields contain comment data that came directly from the customer - check those fields.
Fields that do not include long form textual comment data or other direct customer feedback should not be check-marked. Those fields will be used as metadata filters.
Keep in mind that you need to select at least one text field to be analyzed.
Starting the analysis
Once you've selected the fields to include in the analysis, click the “Review and Upload” box at the bottom of page. Take a second look to make sure everything is correct. Select "upload" and sit back and relax while the Keatext system gets to work!
If you encounter any problems in the upload process, check that: