The CSV format is one of the most common formats to export datasets. A CSV file can be opened on tools such as Excel that provides you with a table view of the data.
On this page, we explain how to upload a CSV file to the Relevance AI platform.
There are special preprocessing that help improve your experience on Relevance AI's platform. For a CSV file, we categorize them under Field/column names and Values.
- Short but descriptive
- Can only contain letters, numbers, dashes or underscores
- All values have a header
- All values under a column follow the same format
- Recommended: For categorical fields (e.g. ”subscriber”), use labels (e.g. “no” / “yes”) instead of numeric codes (e.g. 0 / 1 )
- Recommended: Ensure any date fields are in “yyyy-mm-dd” format
- Recommended: If there is no values in a cell, make sure there is absolutely nothing typed in the cell (i.e. hit delete on the cell) instead of values such as white spaces, none, nil, NA
Sign in to your Relevance AI account, go to the Datasets tab and click on "Create dataset".
A new page similar to the image below with open where you can select the datatype and proceed to upload. By default CSV is selected. However, you can also
- upload image files
- upload audio files
- upload pdf files
- use sample datasets provided by the platform
- automatically get review data from platforms like AppStore
Drag and drop your file(s), type a name for your dataset and hit Next. If you do not type a name for your dataset, the file name is used as the dataset name.
For CSV files, if the file is under the accepted format, a table view of the data will be displayed on the page and the "Upload" button will be activated. Click on the "Upload" to continue or use "Cancel upload" if you need to modify your source CSV file.
Upload time depends on the dataset size but it is normally less than 10 seconds and you will be directed to the dataset when the upload is completed.
If the CSV file is not in the correct format or not following the required conventions, the platform will provide you with an explanation of the issue. Below is one sample:
The following error usually appears when updating a dataset. Imagine an existing dataset contains a field X with only digit or float values. If you try to add entries to the dataset but include any string values under field X in your CSV file, you will be notified of formatting type issue.
Solution is to update the file and keep the data format the same across the dataset.
Updated 1 day ago