Transposit enables developers to create complex SQL-like queries spanning multiple services’ APIs. For example, I can pull records from Airtable and create Stripe charges for each match, all in a single query:
Now you can go even further, using Algorithmia’s Machine Learning APIs to acquire data or transform results.
If you’re a datascientist or developer building your own Machine Learning models, you can use them as well: just deploy your model to Algorithmia, then mix it into a Transposit app!
The actual steps to use Algorithms in Transposit are pretty simple… just use the Algorithmia Connector as a datasource inside your own Transposit App, making sure to set the Authorization to “Simple YOUR_API_KEY” when you add it (where YOUR_API_KEY comes from your Algorithmia account). Then, in the
WHERE clause of your query, set
algorithm to the name of the Algorithm you want to use, and
$body to the JSON input you want to send. Here’s a simple call to nlp/SentimentAnalysis using a fixed input:
Of course, in a real use case you’d probably be pulling many records from another data source, such as the titles of your incoming emails or your Slack messages, then using the resultant sentiment scores to flag particularly high- or low-scoring content. Just for fun, let’s add in the github connector and analyze the sentiment scores of commit messages (looks like seattlerb was having a bad day):
Transposit and Algorithmia are flexible enough to put together just about any workflow you need!
Ready to get started? Start by forking the demo application on Transposit (and click “Connect” under “Auth & user settings” to set “Simple YOUR_API_KEY”), then mix up your own solution: