[{"content":"We are thrilled to announce that PyPOTS has been selected to join the PyTorch Ecosystem! This is a significant milestone for our project, and we are excited to be part of such a vibrant and innovative community.\n🔔 Update on 19 May 2025 PyTorch Ecosystem has been renamed to PyTorch Landscape, and you can find PyPOTS in the PyTorch Landscape.\n","permalink":"https://pypots.com/posts/230623_join_pytorch_eco/","summary":"We are thrilled to announce that PyPOTS has been selected to join the PyTorch Ecosystem! This is a significant milestone for our project, and we are excited to be part of such a vibrant and innovative community.\n🔔 Update on 19 May 2025 PyTorch Ecosystem has been renamed to PyTorch Landscape, and you can find PyPOTS in the PyTorch Landscape.","title":"PyPOTS has been selected to join PyTorch Ecosystem!"},{"content":" Our Vision \"Instead of a flaw to be concealed, the partially observed is a reality to be modeled. PyPOTS exists to provide every practitioner and researcher with the tools to transform incomplete data into comprehensive insights.\" — Wenjie Du, Founder of PyPOTS Real-world time-series data is almost never perfect. Sensors drop out, patients miss hospital visits, satellites lose signal, and yet the patterns hidden inside these partially-observed records hold enormous scientific and practical value. PyPOTS (Python toolbox for Partially-Observed Time Series) was born from a simple conviction: nobody should have to rebuild the same POTS modeling and analysis pipeline from scratch, over and over again. Now PyPOTS has integrated 50+ algorithms, which can end-to-endly handle 5 mainstream time series analysis tasks (imputation, forecasting, classification, clustering, and anomaly detection). 🔬 Research-Grade Quality Every algorithm in PyPOTS ships with peer-reviewed implementations, unified evaluation protocols, and reproducible benchmarks, so your results mean the same thing tomorrow as they do today. ⚡ Practitioner-Friendly API A consistent fit / predict interface across every model means you can swap algorithms in seconds, not days. We obsess over developer experience so you can focus on science. 🌐 Open \u0026amp; For Everyone PyPOTS is BSD-3 licensed and will remain free forever. We believe open science accelerates discovery and that the best tools belong to the whole community. 🏥 Real-World Impact From healthcare monitoring to industrial IoT and climate science, PyPOTS is built for the domains where missing data is not an edge case — it is the norm. Community \u0026amp; Adoption PyPOTS is sustained by a worldwide open-source community of researchers, engineers, and data scientists. The numbers below reflect the collective energy behind the project. 50+ Algorithms Integrated 2M+ PyPI Downloads 2K+ GitHub Stars 20+ Contributors 1K+ Citing Publications Researchers at universities worldwide, spanning North America, Europe, and Asia, rely on PyPOTS to benchmark their models. Industry teams (in healthcare, astronomy, hydrology, geography, manufacturing, and power sector, etc.) deploy it to handle missing sensor streams in production. Every GitHub issue filed, every pull request merged, and every question answered in the community is a brick in this shared foundation. Want to join? We welcome contributions of all sizes, new algorithms, documentation improvements, bug reports, or just a star ⭐ on GitHub. ","permalink":"https://pypots.com/about/","summary":"Our Vision \"Instead of a flaw to be concealed, the partially observed is a reality to be modeled. PyPOTS exists to provide every practitioner and researcher with the tools to transform incomplete data into comprehensive insights.\" — Wenjie Du, Founder of PyPOTS Real-world time-series data is almost never perfect. Sensors drop out, patients miss hospital visits, satellites lose signal, and yet the patterns hidden inside these partially-observed records hold enormous scientific and practical value.","title":"About PyPOTS"},{"content":" Research institutes: Industry companies: Note: only followers, community members, and people citing PyPOTS or reaching out to our team are listed as users here. Stargazers of PyPOTS repos are not included. Didn't see your affiliation in the list? Drop us an email telling how you're using PyPOTS to add your affiliation above. ","permalink":"https://pypots.com/users/","summary":" Research institutes: Industry companies: Note: only followers, community members, and people citing PyPOTS or reaching out to our team are listed as users here. Stargazers of PyPOTS repos are not included. Didn't see your affiliation in the list? Drop us an email telling how you're using PyPOTS to add your affiliation above. ","title":"Affiliations of PyPOTS Users"},{"content":" In PyPOTS, everything is part of a coffee universe. The coffee pot in our logo is just the beginning. Follow the pipeline below to see how each library contributes to end-to-end research and practice on partially-observed time series. TSDB A dataset hub for open time-series benchmarks. In the PyPOTS universe, datasets are coffee beans, and POTS datasets are beans with meaningful missing parts. TSDB (Time Series Data Beans) makes these resources easy to access through a unified loading interface for a wide range of public datasets. It currently supports 173 open-source datasets, helping you start experiments quickly and reproducibly. PyGrinder A toolkit for generating realistic missingness in time series. PyGrinder is built to simulate real-world incompleteness by \"grinding\" full datasets into partially-observed ones. It supports all major missingness mechanisms based on Rubin's framework: MCAR, MAR, and MNAR. With simple APIs, you can inject synthetic missing values and control missing patterns for robust model evaluation. BenchPOTS A unified benchmarking suite for machine learning on POTS. BenchPOTS provides standardized preprocessing and evaluation pipelines for partially-observed time-series tasks. It helps researchers and engineers compare methods fairly across datasets, settings, and task types. With consistent protocols, benchmarking becomes more transparent, reproducible, and decision-friendly. PyPOTS The core toolbox for modeling partially-observed time series. Once the beans are prepared, PyPOTS is the coffee pot that brews them into high-value analysis results. PyPOTS integrates dozens of algorithms with unified APIs for end-to-end imputation, forecasting, classification, clustering, and anomaly detection. Since 2022, it has been widely adopted in scientific and industrial projects, with growing citations and references. Here is an incomplete list. BrewPOTS The tutorial repository for practical PyPOTS workflows. With the beans, grinder, benchmark, and pot in place, BrewPOTS shows you how to put everything together. It hosts practical tutorials, examples, and walkthroughs so you can move from setup to results efficiently. Visit BrewPOTS to learn how to build complete POTS workflows in real projects. ☕️ Welcome to the PyPOTS universe. ","permalink":"https://pypots.com/ecosystem/","summary":"In PyPOTS, everything is part of a coffee universe. The coffee pot in our logo is just the beginning. Follow the pipeline below to see how each library contributes to end-to-end research and practice on partially-observed time series. TSDB A dataset hub for open time-series benchmarks. In the PyPOTS universe, datasets are coffee beans, and POTS datasets are beans with meaningful missing parts. TSDB (Time Series Data Beans) makes these resources easy to access through a unified loading interface for a wide range of public datasets.","title":"Ecosystem"},{"content":" [📜 Paper] Fengming Zhang, Wenjie Du, Huan Zhang, Ke Yu, Shen Qu. HELIX: Hybrid Encoding with Learnable Identity and Cross-dimensional Synthesis for Time Series Imputation, ICML (spotlight), 2026. [DOI Link] [PDF] [Code] [📜 Paper] Wenjie Du, Yiyuan Yang, Tianxiang Zhan, Qingsong Wen. End-to-End Learning for Partially-Observed Time Series with PyPOTS, KDD, 2026. [DOI Link] [PDF] [📜 Paper] Tianxiang Zhan, Yuanpeng He, Yong Deng, Zhen Li, Wenjie Du, Qingsong Wen. Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting, TPAMI, 2025. [DOI Link] [PDF] [Code] 2024 IF-18.6/JCR-Q1/CAS-Q1 [🎙📚️ Tutorial] Wenjie Du, Yiyuan Yang, Linglong Qian. PyPOTS \u0026amp; Datawhale Team Learning, May 2025. (Datawhale is the largest open-source community in China, and we are honored to be invited to give a tutorial for their learners on time series analysis with PyPOTS.) [Call for Learners] [Tutorial] [📜 Paper] Jun Wang*, Wenjie Du*, Yiyuan Yang, Linglong Qian, Wei Cao, Keli Zhang, Wenjia Wang, Yuxuan Liang, Qingsong Wen. Deep Learning for Multivariate Time Series Imputation: A Survey, IJCAI, 2025. [DOI Link] [PDF] [Code] [📜 Paper] Yaxuan Kong*, Yiyuan Yang*, Yoontae Hwang, Wenjie Du, Stefan Zohren, Zhangyang Wang, Ming Jin, Qingsong Wen. Time-MQA: Time Series Multi-Task Question Answering with Context Enhancement, ACL (Main Conference), 2025. [DOI Link] [PDF] [Code] [🎙️ Talk] Wenjie Du. Learning from POTS: Towards Reality-Centric AI for Time Series, Keynote at IJCAI'24 AI4TS (AI for Time Series Analysis) Workshop, Jeju, South Korea, August 5, 2024. [Slides] [📝 Preprint] Wenjie Du, Jun Wang, Linglong Qian, Yiyuan Yang, Zina Ibrahim, Fanxing Liu, Zepu Wang, Haoxin Liu, Zhiyuan Zhao, Yingjie Zhou, Wenjia Wang, Kaize Ding, Yuxuan Liang, B. Aditya Prakash, Qingsong Wen. TSI-Bench: Benchmarking Time Series Imputation, arXiv preprint, abs/2406.12747, 2024. [DOI Link] [PDF] [Code] [📜 Paper] Linglong Qian, Zina Ibrahim, Wenjie Du, Yiyuan Yang, Richard JB Dobson. Unveiling the Secrets: How Masking Strategies Shape Time Series Imputation, IJCAI'24 AI4TS (AI for Time Series Analysis) Workshop, 2024. [DOI Link] [PDF] [Code] [📜 Paper] Wenjie Du. PyPOTS: A Python Toolbox for Data Mining on Partially-Observed Time Series, KDD MiLeTS Workshop, 2023. [PDF] [Code] [📜 Paper] Wenjie Du, David Cote, Yan Liu. SAITS: Self-Attention-based Imputation for Time Series, Expert Systems with Applications, 219:119619, 2023. [DOI Link] [PDF] [Code] 2022 IF-8.665/JCR-Q1/CAS-Q1 \u0026nbsp; [📜 Paper] Wenjie Du, David Cote, Chris Barber, Yan Liu. Forecasting Loss of Signal in Optical Networks with Machine Learning, IEEE/OSA Journal of Optical Communications and Networking, vol. 13, no. 10, pp. E109-E121 (2021). [DOI Link] [PDF] [Code] 2022 IF-4.508/JCR-Q1/CAS-Q1 [🎙️ Talk] Wenjie Du. Efficient and Effective Time Series Imputation: from RNN to Transformer, Virtual Presentation at Ciena Corp., Montreal, Canada, July 9, 2021. ","permalink":"https://pypots.com/pubs/","summary":"[📜 Paper] Fengming Zhang, Wenjie Du, Huan Zhang, Ke Yu, Shen Qu. HELIX: Hybrid Encoding with Learnable Identity and Cross-dimensional Synthesis for Time Series Imputation, ICML (spotlight), 2026. [DOI Link] [PDF] [Code] [📜 Paper] Wenjie Du, Yiyuan Yang, Tianxiang Zhan, Qingsong Wen. End-to-End Learning for Partially-Observed Time Series with PyPOTS, KDD, 2026. [DOI Link] [PDF] [📜 Paper] Tianxiang Zhan, Yuanpeng He, Yong Deng, Zhen Li, Wenjie Du, Qingsong Wen. Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting, TPAMI, 2025.","title":"Papers and Talks from PyPOTS"}]