Mozilla Data Collective Alpha Goes Live

Mozilla Data Collective is live in Alpha – the new platform from Mozilla Foundation that puts communities in control of how datasets are shared! Starting today, the latest Common Voice datasets (23.0) are available to download through Mozilla Data Collective. We’re so excited to introduce 149 new languages in this release, alongside an important first: Spontaneous Speech datasets, which include transcribed, spontaneous responses to prompts that help train models on more realistic speech patterns. 

Create a Mozilla Data Collective account to download all the datasets. From there, you can use the new datasheets pages to explore and access resources, access. You can also get access via API or integrate them easily with our new open-source Python library, which allows easy access to datasets programmatically.  This means increased global access, more options for dataset downloads and new ways to work with data at scale. 

Common Voice datasets are just the beginning! Mozilla Data Collective is a platform to let all dataset communities and creators share their data under their own terms. We’ll be adding more partner datasets soon. If you have a dataset you would like to see on Mozilla Data Collective, tell us about it at mozilladatacollective@mozillafoundation.org.

Mozilla Data Collective wouldn’t exist without the language communities, dataset users and amazing community that makes up Common Voice. Thank you for building with us. We’re excited to hear your feedback, wishlists and ideas for what you want us to build, so get in touch at mozilladatacollective@mozillafoundation.org.