Our clients have been using Brandwatch for a multitude of different use cases. Here are some pointers for how to get the best performance when integrating the Analytics API with your application.
API Clients are limited to make 30 calls every 10 minutes. Limits are applied at a Client rather than User level. If you have a use case that doesn't fit within these rate limits, then please get in touch with your account manager to discuss your needs.
If you are building an integration that uses polling in order to keep a dataset continually up to date, think about the type of data that you are displaying in order to choose appropriate polling times.
For example, if you were populating a stream of Mentions, then it would make sense to poll fairly frequently (e.g. once every 30 seconds) as new data will be arriving continually. However, if you were populating a chart which breaks down volume by week, then choosing a longer polling time would be more appropriate as the data will not change as significantly.
We recommend queuing requests at the client side and executing them linearly, rather than in parallel. Multiple parallel requests may be throttled and take longer to return than a sequence of linearly executed requests.
If you are building an application, we recommend caching Categories, Tags, and Lists locally, so that you do not have to retrieve them every time that you do a Mentions data call.
If you're looking to quickly get started with experimenting with our API, there is a Python utility library that has is developed and maintained by our Professional Services team. It's a continual work-in-progress, so please submit pull requests if you have any improvements.