Connecting data to DeepSea allows users to view & explore the data, create visualizations using it, and use it as input to a python function.

There are several ways to ingest or create data in DeepSea: uploading a CSV file, creating a model that outputs data, or creating a form to collect data.

This article will focus on connecting DeepSea to your existing data in a database, data lake, or file.

Uploading CSV Files

The easiest way to get started with data in DeepSea is to import a CSV file. For example, many CSV files are available on Kaggle. Or, you may have an Excel file of data that you wish to import: to do this, you can export the sheet as a .csv file from Excel and then import it into DeepSea. It is best to make sure that your CSV has column names as its first row.

To upload a CSV file to DeepSea, navigate to the Datasets page in a project and click the “Add” button next to “Datasets”.

Then, choose “Upload File”. You can drag the file into the box, or click the button to choose the file from your computer. You can give the dataset a name, or it will default to the name of the file - you can always change this later. 

Click “upload” and your upload’s progress will appear in the bottom left sidebar. It may take a few minutes to upload if the file is large.

Once a green checkmark appears, your data is ready to be viewed in DeepSea by clicking the name of the dataset in the sidebar. It can now be used in dashboards or models.

CSV Upload Troubleshooting

If the upload or ingestion fails, a red icon will appear with an error message. This is usually due to something in the data that our SQL database cannot handle, such as mismatched quote characters around strings, varying numbers of columns, or a file or character that is not readable in the UTF-8 format.

If this happens, it is good to check that the data looks right and has no issues being opened in other programs such as Excel or Google Sheets, or by Pandas in Python.

Connecting to a Database

Currently, the easiest way to connect to a database or data lake for data that updates periodically is to write a python script that pulls the data using an API or database driver, and then schedule that script to run regularly.

Please contact our team for more help setting this up.

If your database data is static, it is easiest to export it as a CSV and then follow the CSV file ingestion steps.

Working With Features

Whenever table data is imported into DeepSea from any source, DeepSea will automatically create features which represent the columns of the table, allowing you to control their names and types.

There are currently four types of features in DeepSea:

  • number: any number column (either integers or floats)
  • date: columns containing dates and/or times
  • category: columns containing text that falls into one of a few categories
  • text: columns containing text that is not categorical or is more detailed

These feature types allow DeepSea to display and query the data correctly. 

The features also have display names - what will display to the end user when viewing data or visualizations in DeepSea, and SQL column names - the name used for the column in the CSV, database, and/or python script.

Up Next:

5. Build Your First Dashboard

Learn how to create a dashboard and add/edit the visualizations within it.