Here are several example use cases for Mage (the list is not comprehensive):
Use cases meme
Data integration is synchronizing data from an external system to an internal system or vice versa.
If you have data in 3rd party systems (e.g other SaaS tools) and you want that data available in your database, data warehouse, or data lake, you can use Mage to sync data between those systems.
Your internal database or data warehouse may contain application data generated by your users or customers. This data can be valuable if used in conjunction with 3rd party SaaS tools.
You can use Mage to sync data from your internal systems, and have that data appear in various 3rd party SaaS tools.
Sync user event data from Amplitude and save it in a table in a Snowflake database.
Sync booking transaction data from PostgreSQL and save it in Salesforce.
- Facebook Ads
- Google Ads
- Google Analytics
- Google Sheets
- and more...
- AWS DynamoDB
- AWS Redshift
- AWS S3
- Azure Blob Storage
- Azure Synapse Analytics
- GCP BigQuery
- GCP Cloud Storage
- Oracle Database
- and more...
Once data is in your internal systems (e.g. database, data warehouse, data lake, etc.), you can write code to clean the data, reformat the data, enhance the data, enrich the data, fill in missing data, aggregate the data, and much more. These are all examples of transforming your data.
During data analysis, many analysts or data scientists transform raw data to build insights or visualizations to better understand the shape and trend of data.
- Combine user demographics and website activity data and see which users from which country is more engaged with the product.
- Aggregate purchase history data, group each row by date, and sum the values in the revenue column to create a time-series line chart showing total revenue by day.
Once data is synced, stored, and analyzed, it can then be transformed further to make predictions.
Machine learning (ML) models are good at learning from past data to make future predictions.
Before training a machine learning model, a process called data preparation and feature extraction is required to build a training set that the model learns from.
That process is iterative and requires many data transformations to produce an effective training set that’ll maximize the machine learning model’s performance.
- Group user click data by user ID, then create a column that counts number of clicks for that user over 3 days, 7 days, and 14 days. This produces 3 new columns that the ML model can learn from.
- Extract the area code integers from a user’s phone number, find the corresponding city, and create a column from that value.