​From social networks to the Internet of Things, the number of applications that generate huge quantities of data has grown exponentially over the past decade. This has created a need for datastream processing software with machine learning capabilities to mine data for information and extract value. CEA-List's open-source Streamer environment was designed for datastream processing and analysis software developers and scientists developing the algorithms that power this kind of software.

Datastream processing software has to be able to handle continuous datastreams without storing the data. To do so, the software must be able to learn and adapt in real time. And for developers to evaluate datastream software, they need to be able to test it in realistic datastreaming scenarios. Streamer provides this realistic testing environment by simulating different data transmission and receival parameters for different operating contexts. Classification algorithms, neural networks, and other pre- and post-processing functions are integrated into the simulator. Users can add new Python or R algorithms easily. Streamer also includes different metrics for evaluating algorithms, so that users can easily test them during the development process.

Streamer is fast and easy to install and is compatible with all operating systems (Windows, Linux, MacOS). The graphical user interface was designed to deliver a no-code testing experience. The delivery model is open source, with a community of developers constantly improving and updating the tool.

*Open-source code available (https://streamer-framework.github.io/) under GPLv3 license. Official website: https://streamer-framework.github.io.