Creating a data connection in Pecan enables you to import data from a data source (“Read” connection) or export predictions to a data destination (“Write” connection). Pecan offers a variety of connectors so you can easily connect multiple data sources/destinations to the platform, as per your company’s data governance policy.
This article gives an overview of the two different ways you can connect data to the platform: via hosted files and via a data service. (As a simple alternative for importing data, Pecan also lets you upload local CSV files.)
It then covers the three main steps for creating a data connection:
Connecting to hosted files
The simplest way to connect your data is via an Amazon S3 bucket. This is optimal if your database can’t be accessed through SaaS, or if you use a dedicated standalone service. Since this method requires the development of an ETL service, you should consider the long-term effort of maintaining it.
Pecan supports read/write connections for the following file formats:
CSV files
Delta tables
Parquet and Delta files are compressed and organized by column rather than row, which saves storage space, speeds up analytics queries, and eliminates the need for manual data-types review.
Connecting to a data service
Establishing a direct connection between Pecan and your database, server or warehouse enables continuous data collection with no requirement for ETL services. You may grant Pecan permissions according to your company’s security policy.
Pecan supports the following cloud connections:
Data service | Connection Type |
| Data service | Connection Type |
Adjust | Write |
| Read/Write | |
Appsflyer | Write |
| Read/Write | |
Read/Write |
| Read/Write | ||
Read/Write |
| Read/Write | ||
Databricks |
|
| Read/Write | |
Firebase | Write |
| Salesforce | Read/Write |
Read/Write |
| Singular | Write | |
Read |
| Read/Write |
More connectors are being added on a continual basis.
How to create a data connection
Whether connecting a data service or S3-hosted files to Pecan, the procedure is similar.
1. Go to the Connections tab
Log in to Pecan and go to the “Connections” tab.
If you are inside a Nutbook you can click the "Add data" button in the data panel:
Click + New connection.
2. Select a connection type
You'll see all available data sources/destinations – select the relevant one.
3. Configure the connection
Now that you’ve selected a data source/destination, you’ll need to configure the connection.
For step-by-step instructions on how to configure each connection type, follow the above links for hosted files and data services. You'll also find useful information on the right side of the connection creation page.
4. Test and finalize the connection
Now that you’ve configured the connection, it’s time to make sure it works – and then finalize the creation process.
While still on the configuration screen, click Test connection. This process may take anywhere from a few seconds to a couple of minutes, depending on the size of the database.
If the test is successful, click Create.
If the test fails, check out this troubleshooting article and consult with your internal IT team to ensure you have all the proper credentials. If that doesn’t help, feel free to contact Pecan support.
That’s it! Your data connection has been created. You can now import the relevant data to Pecan so it can be used to train models and generate predictions.
The new connection will now appear on the left side of the “Connections” screen, as illustrated below:
You can use the “Configuration” tab to view, edit, or delete the connection; use the “Tables” tab to view its constituent tables (for “Read” connections); and use the “History” tab to view import and export jobs for the connection.
What if my data service isn't natively supported by Pecan?
You can easily utilize third-party services like Airbyte, Fivetran, or Rivery (and more) to convert your data into a Parquet file, then store it in an S3 bucket.
Connect Pecan to this bucket to both automate fresh data flow for predictions and effortlessly route those predictions back to your warehouse. This flexible workaround ensures you can still harness the full power of Pecan's predictive analytics.
For more detailed instructions, please refer to our guide on creating an S3 data connection.