```演进中的 FSQ 开源场所 (2025)```
Evolving FSQ Open Source Places (2025)

原始链接: https://foursquare.com/resources/blog/data/evolving-fsq-open-source-places/

自 2024 年开源 Places 数据集以来,Foursquare 在其社区驱动的模式下取得了显著成功,编辑次数超过 2700 万次,下载量也实现了大幅增长。为了进一步推进这一进展,Foursquare 正在改进用户访问数据的方式,以促进数据集使用者与“Placemaker”社区之间建立更紧密的联系。 自 2025 年 10 月起,Foursquare 将从公共 S3 访问转向全新的 **Places Portal**,用户需注册以生成访问令牌。在保持适当署名的前提下,该数据集将继续根据 Apache 2.0 许可证免费提供。数据也将继续通过 Snowflake Marketplace 和 HuggingFace 提供访问。 此次转变旨在弥合应用程序开发者与维护数据的贡献者之间的鸿沟。通过与使用者建立直接关系,Foursquare 可以更好地将社区工作与现实需求相结合,例如直接向终端用户呈现位置更新或企业停业信息。这种“良性循环”鼓励更多用户成为贡献者,从而确保数据集的准确性和可持续性。建议现有基于 S3 工作负载的开发者联系 Foursquare 以获取迁移支持。

Hacker News 最新 | 过往 | 评论 | 提问 | 展示 | 招聘 | 提交 登录 演进中的 FSQ 开源场所数据 (2025) (foursquare.com) 3 点,由 altilunium 发布于 1 小时前 | 隐藏 | 过往 | 收藏 | 讨论 | 帮助 指导方针 | 常见问题 | 列表 | API | 安全 | 法律 | 加入 YC | 联系 搜索:
相关文章

原文

In November 2024, we took a bold step in open-sourcing our Places dataset with a hypothesis that a community driven approach is the only way to create a sustainable and robust places dataset that can serve the needs of diverse problem domains. Our thesis has been validated with the tremendous progress we have been able to make in improving this open dataset since its inception.

  1. Increased adoption: Our S3 listing is downloaded from more than 5000 unique IPs every month, a steady increase from 500 unique IPs when we first launched. More than 250k queries are executed on our Snowflake listing every month. And we have 3000+ monthly downloads of our dataset on HuggingFace.
  2. Accelerated data improvements: We continuously refined our dataset adding more than a million places since launch with a monthly high of 160k+ new places just in September. More than 27 million edits have been proposed by our Placemaker community in the last year out of which close to 17 million have been resolved.
  3. Diversified Placemaker community: We have been able to diversify our community from just the users of our apps to a broader user base with around 2000 Placemaker sign-ups from the open-source community using our Placemaker tools. We have also seen business owners gravitating towards our Placemaker Tools to update their own listings.

Today, we’re taking the next step in that journey with updates that will strengthen our collaborative ecosystem and unlock even greater potential for data improvement.

The Power of Placemaker Tools

At the heart of our community-driven approach are our Placemaker Tools, which enable anyone to contribute to and improve the OS Places dataset. Through these tools, contributors can:

  • Add new places: Submit businesses, landmarks, and points of interest missing from the dataset
  • Update existing information: update addresses, phone numbers, business hours, and other details
  • Verify and validate: Confirm the accuracy of place information through community consensus
  • Enhance place details: Add categories, attributes, and rich contextual information
  • Report closures: Flag businesses that have permanently closed or relocated

These tools have proven that when given accessible ways to contribute, people eagerly improve the data that powers their favorite applications. The 27 million edits proposed through Placemaker Tools demonstrate the power of community-driven data maintenance at scale. As we evolve our approach, Placemaker Tools remain central to our strategy. By creating stronger connections between dataset consumers and the improvement ecosystem, we can channel community efforts where they matter most, whether that’s improving coverage in underserved regions, updating time-sensitive information, or enriching place details for specific use cases.

Unveiling the Next Chapter

Our October release transitions FSQ OS Places from public S3 bucket access to our new Places portal. Users can register, generate an access token, and retrieve the data through an Iceberg catalog. The data remains completely free with proper attribution under our Apache 2.0 license.

Why This Evolution Matters

Over the past year, we’ve seen incredible adoption of our dataset across diverse industries and use cases. But despite the widespread adoption of our OS Places dataset, the end users of applications built on this dataset rarely know it is even sourced from Foursquare or that they have an opportunity to improve it using our Placemaker Tools. In our current anonymous distribution model, there’s no path from “using an app with great location data” to “helping improve that location data.”

A root cause here is a lack of awareness. When users discover that their favorite applications are powered by a community-driven dataset, many become eager contributors through our Placemaker Tools. This creates a virtuous cycle: better data enables better applications, which reach more users, who contribute more improvements, creating even better data.

That is why we’re introducing this new approach to build connections with the direct consumers of our dataset and work together to spread awareness of FSQ OS Places and empower their user communities to contribute through Placemaker Tools.

What This Means for You

Accessing the dataset

FSQ OS Places will now be accessible through three primary channels:

  1. New Places Portal: Visit our new portal to create your free account and generate an access token. Use the access token to retrieve OS Places data from our Iceberg catalog.
  2. Snowflake: Access OS Places data from the Snowflake marketplace with the same ease as before.
  3. HuggingFace: Get approved for access to the OS Places data through HuggingFace by providing your contact information.

Attributing to Foursquare

The data remains completely free with proper attribution under the Apache 2.0 license. To ensure required attribution, we recommend the following:

  • If using/distributing the dataset in flat file form as-is or after making changes/modifications: include this NOTICE.txt file, which may be modified to include an additional notice of your changes/modifications, if any.
  • If using/distributing the data in API form as-is or after making changes/modifications: include a copy of the content from this NOTICE.txt file prominently in your developer documentation for such API, which may be modified to include an additional notice of your changes/modifications, if any.

This new approach creates direct connections between you, the dataset consumers, and our Placemaker community. When we know which applications are powered by FSQ OS Places, we can help you surface improvement opportunities to your end users, empowering them to become contributors who enhance the data that powers their favorite applications.

When we understand what matters to you — accurate business closures, comprehensive category coverage, and more — we can channel the community’s efforts where they will have the greatest impact. Your needs inform community priorities, their contributions power your applications, and your users become their fellow contributors. The cycle strengthens itself. Join our Placemaker Discord to connect directly with the community and collaborate on data improvements.

Transition Support

We are committed to ensuring a smooth transition for everyone. You can access the OS Places data from the past releases through our public S3 bucket but all future releases (starting from October 2025) will only be accessible through the three supported channels.

The September 2025 release is already available through our new portal, giving you immediate access to our latest data while you plan your transition. If you have production workloads built on top of our public S3 dataset, please join our open source slack and reach out to us at [email protected] and we will work with you to ensure continuity of business operations.

The Road Ahead

This next chapter strengthens everything that makes FSQ OS Places powerful: community collaboration, open access, and continuous improvement. By understanding real-world usage patterns, we can identify data gaps and connect users directly with Placemaker Tools. Every application becomes a gateway for new contributors, and every contribution amplifies the value for all users.

Together, we’re building not just a dataset, but a sustainable ecosystem where location data gets better every day through the collective efforts of a global community.

Visit the new Places Portal to get started.

联系我们 contact @ memedata.com