[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-2471":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":30,"readmeContent":31,"aiSummary":32,"trendingCount":16,"starSnapshotCount":16,"syncStatus":33,"lastSyncTime":34,"discoverSource":35},2471,"prophet","facebook\u002Fprophet","facebook","Tool for producing high quality forecasts for time series data that has multiple seasonality with linear or non-linear growth.","https:\u002F\u002Ffacebook.github.io\u002Fprophet",null,"Python",20223,4630,443,452,0,1,11,52,3,45,"MIT License",false,"main",true,[27,28,29],"forecasting","python","r","2026-06-12 02:00:41","\n# Prophet: Automatic Forecasting Procedure\n\n![Build](https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet\u002Fworkflows\u002FBuild\u002Fbadge.svg)\n\n[![PyPI Version](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fprophet.svg)](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fprophet)\n[![PyPI Downloads Monthly](https:\u002F\u002Fpepy.tech\u002Fbadge\u002Fprophet\u002Fmonth)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fprophet)\n[![PyPI Downloads All](https:\u002F\u002Fpepy.tech\u002Fbadge\u002Fprophet)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fprophet)\n\n[![CRAN Version](https:\u002F\u002Fwww.r-pkg.org\u002Fbadges\u002Fversion\u002Fprophet)](https:\u002F\u002FCRAN.R-project.org\u002Fpackage=prophet)\n[![CRAN Downloads Monthly](https:\u002F\u002Fcranlogs.r-pkg.org\u002Fbadges\u002Fprophet?color=brightgreen)](https:\u002F\u002Fcran.r-project.org\u002Fpackage=prophet)\n[![CRAN Downloads All](https:\u002F\u002Fcranlogs.r-pkg.org\u002Fbadges\u002Fgrand-total\u002Fprophet?color=brightgreen)](https:\u002F\u002Fcranlogs.r-pkg.org\u002Fbadges\u002Fgrand-total\u002Fprophet)\n\n[![Conda_Version](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fprophet\u002Fbadges\u002Fversion.svg)](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fprophet\u002F)\n\n-----\n\n**2023 Update:** We discuss our plans for the future of Prophet in this blog post: [facebook\u002Fprophet in 2023 and beyond](https:\u002F\u002Fmedium.com\u002F@cuongduong_35162\u002Ffacebook-prophet-in-2023-and-beyond-c5086151c138)\n\n-----\n\nProphet is a procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects. It works best with time series that have strong seasonal effects and several seasons of historical data. Prophet is robust to missing data and shifts in the trend, and typically handles outliers well.\n\nProphet is [open source software](https:\u002F\u002Fcode.facebook.com\u002Fprojects\u002F) released by Facebook's [Core Data Science team](https:\u002F\u002Fresearch.fb.com\u002Fcategory\u002Fdata-science\u002F). It is available for download on [CRAN](https:\u002F\u002Fcran.r-project.org\u002Fpackage=prophet) and [PyPI](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fprophet\u002F).\n\n## Important links\n\n- Homepage: https:\u002F\u002Ffacebook.github.io\u002Fprophet\u002F\n- HTML documentation: https:\u002F\u002Ffacebook.github.io\u002Fprophet\u002Fdocs\u002Fquick_start.html\n- Issue tracker: https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet\u002Fissues\n- Source code repository: https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet\n- Contributing: https:\u002F\u002Ffacebook.github.io\u002Fprophet\u002Fdocs\u002Fcontributing.html\n- Prophet R package: https:\u002F\u002Fcran.r-project.org\u002Fpackage=prophet\n- Prophet Python package: https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fprophet\u002F\n- Release blogpost: https:\u002F\u002Fresearch.facebook.com\u002Fblog\u002F2017\u002F2\u002Fprophet-forecasting-at-scale\u002F\n- Prophet paper: Sean J. Taylor, Benjamin Letham (2018) Forecasting at scale. The American Statistician 72(1):37-45 (https:\u002F\u002Fpeerj.com\u002Fpreprints\u002F3190.pdf).\n\n## Installation in R - CRAN\n\n⚠️ **The CRAN version of prophet is fairly outdated. To get the latest bug fixes and updated country holiday data, we suggest installing the [latest release](#installation-in-r---latest-release).**\n\nProphet is a [CRAN package](https:\u002F\u002Fcran.r-project.org\u002Fpackage=prophet) so you can use `install.packages`.\n\n```r\ninstall.packages('prophet')\n```\n\nAfter installation, you can [get started!](https:\u002F\u002Ffacebook.github.io\u002Fprophet\u002Fdocs\u002Fquick_start.html#r-api)\n\n## Installation in R - Latest release\n\n```r\ninstall.packages('remotes')\nremotes::install_github('facebook\u002Fprophet@*release', subdir = 'R')\n```\n\n#### Experimental backend - cmdstanr\n\nYou can also choose an experimental alternative stan backend called `cmdstanr`. Once you've installed `prophet`,\nfollow these instructions to use `cmdstanr` instead of `rstan` as the backend:\n\n```r\n# R\n# We recommend running this in a fresh R session or restarting your current session\ninstall.packages(c(\"cmdstanr\", \"posterior\"), repos = c(\"https:\u002F\u002Fstan-dev.r-universe.dev\", getOption(\"repos\")))\n\n# If you haven't installed cmdstan before, run:\ncmdstanr::install_cmdstan()\n# Otherwise, you can point cmdstanr to your cmdstan path:\ncmdstanr::set_cmdstan_path(path = \u003Cyour existing cmdstan>)\n\n# Set the R_STAN_BACKEND environment variable\nSys.setenv(R_STAN_BACKEND = \"CMDSTANR\")\n```\n\n### Windows\n\nOn Windows, R requires a compiler so you'll need to [follow the instructions](https:\u002F\u002Fgithub.com\u002Fstan-dev\u002Frstan\u002Fwiki\u002FRStan-Getting-Started) provided by `rstan`. The key step is installing [Rtools](http:\u002F\u002Fcran.r-project.org\u002Fbin\u002Fwindows\u002FRtools\u002F) before attempting to install the package.\n\nIf you have custom Stan compiler settings, install from source rather than the CRAN binary.\n\n## Installation in Python - PyPI release\n\nProphet is on PyPI, so you can use `pip` to install it.\n\n```bash\npython -m pip install prophet\n```\n\n* From v0.6 onwards, Python 2 is no longer supported.\n* As of v1.0, the package name on PyPI is \"prophet\"; prior to v1.0 it was \"fbprophet\".\n* As of v1.1, the minimum supported Python version is 3.7.\n\nAfter installation, you can [get started!](https:\u002F\u002Ffacebook.github.io\u002Fprophet\u002Fdocs\u002Fquick_start.html#python-api)\n\n### Anaconda\n\nProphet can also be installed through conda-forge.\n\n```bash\nconda install -c conda-forge prophet\n```\n\n## Installation in Python - Development version\n\nTo get the latest code changes as they are merged, you can clone this repo and build from source manually. This is **not** guaranteed to be stable.\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet.git\ncd prophet\u002Fpython\npython -m pip install -e .\n```\n\nBy default, Prophet will use a fixed version of `cmdstan` (downloading and installing it if necessary) to compile the model executables. If this is undesired and you would like to use your own existing `cmdstan` installation, you can set the environment variable `PROPHET_REPACKAGE_CMDSTAN` to `False`:\n\n```bash\nexport PROPHET_REPACKAGE_CMDSTAN=False; python -m pip install -e .\n```\n\n### Linux\n\nMake sure compilers (gcc, g++, build-essential) and Python development tools (python-dev, python3-dev) are installed. In Red Hat systems, install the packages gcc64 and gcc64-c++. If you are using a VM, be aware that you will need at least 4GB of memory to install prophet, and at least 2GB of memory to use prophet.\n\n### Windows\n\nUsing `cmdstanpy` with Windows requires a Unix-compatible C compiler such as mingw-gcc. If cmdstanpy is installed first, one can be installed via the `cmdstanpy.install_cxx_toolchain` command.\n\n## Changelog\n\nSee [Release Notes](https:\u002F\u002Fgithub.com\u002Ffacebook\u002Fprophet\u002Freleases).\n\n### Version 1.3.0 (2026.01.27)\n\n#### Python\n\n- Support pandas>=3.0 and numpy>=2.4.\n\n### Version 1.2.2 (2026.01.25)\n\n#### Python\n\n- Version constraints on pandas (`\u003C3`) and numpy (`\u003C2.4`).\n\n#### R\n- Update build requirements to C++17 to Comply with CRAN Policy.\n- Add .tar.gz upload for R package to CI.\n- Re-generated holidays.csv for R package.\n\n### Version 1.2.1 (2025.10.22)\n\n#### Python\n\n- Also copy makefile to fake cmdstan.\n\n### Version 1.2.0 (2025.05.30)\n\n#### Python\n\n- Use latest CmdStan.\n- Add null check to CmdStanPyBackend cleanup() function.\n\n### Version 1.1.7 (2025.05.30)\n\n#### Python\n\n- Enable creation of custom performance metrics.\n- chore: address pandas futurewarning from \"M\" being deprecated.\n- cleanup() for cross_validate.\n\n### Version 1.1.6 (2024.09.29)\n\n#### Python\n\n- Bug fixes: include predictions for dates with missing `y` the history, zero division error in cross validation metrics.\n- Changed `NDArray[np.float_]` to `NDArray[np.float64]` to be compatible with numpy 2.0\n\n#### R\n\n- Updated `holidays` data based on holidays version 0.57.\n\n### Version 1.1.5 (2023.10.10)\n\n#### Python\n\n- Upgraded cmdstan version to 2.33.1, enabling Apple M2 support.\n- Added pre-built wheels for macOS arm64 architecture (M1, M2 chips)\n- Added argument `scaling` to the `Prophet()` instantiation. Allows `minmax` scaling on `y` instead of\n  `absmax` scaling (dividing by the maximum value). `scaling='absmax'` by default, preserving the\n  behaviour of previous versions.\n- Added argument `holidays_mode` to the `Prophet()` instantiation. Allows holidays regressors to have\n  a different mode than seasonality regressors. `holidays_mode` takes the same value as `seasonality_mode`\n  if not specified, preserving the behaviour of previous versions.\n- Added two methods to the `Prophet` object: `preprocess()` and `calculate_initial_params()`. These\n  do not need to be called and will not change the model fitting process. Their purpose is to provide\n  clarity on the pre-processing steps taken (`y` scaling, creating fourier series, regressor scaling,\n  setting changepoints, etc.) before the data is passed to the stan model.\n- Added argument `extra_output_columns` to `cross_validation()`. The user can specify additional columns\n  from `predict()` to include in the final output alongside `ds` and `yhat`, for example `extra_output_columns=['trend']`.\n- prophet's custom `hdays` module was deprecated last version and is now removed.\n\n#### R\n\n- Updated `holidays` data based on holidays version 0.34.\n\n### Version 1.1.4 (2023.05.30)\n\n#### Python\n\n- We now rely solely on `holidays` package for country holidays.\n- Upgraded cmdstan version to 2.31.0, enabling Apple M1 support.\n- Fixed bug with Windows installation caused by long paths.\n\n#### R\n\n- Updated `holidays` data based on holidays version 0.25.\n\n### Version 1.1.2 (2023.01.20)\n\n#### Python\n\n- Sped up `.predict()` by up to 10x by removing intermediate DataFrame creations.\n- Sped up fourier series generation, leading to at least 1.5x speed improvement for `train()` and `predict()` pipelines.\n- Fixed bug in how warm start values were being read.\n- Wheels are now version-agnostic.\n\n#### R\n\n- Fixed a bug in `construct_holiday_dataframe()`\n- Updated `holidays` data based on holidays version 0.18.\n\n### Version 1.1.1 (2022.09.08)\n\n- (Python) Improved runtime (3-7x) of uncertainty predictions via vectorization.\n- Bugfixes relating to Python package versions and R holiday objects.\n\n### Version 1.1 (2022.06.25)\n\n- Replaced `pystan2` dependency with `cmdstan` + `cmdstanpy`.\n- Pre-packaged model binaries for Python package, uploaded binary distributions to PyPI.\n- Improvements in the `stan` model code, cross-validation metric calculations, holidays.\n\n### Version 1.0 (2021.03.28)\n\n- Python package name changed from fbprophet to prophet\n- Fixed R Windows build issues to get latest version back on CRAN\n- Improvements in serialization, holidays, and R timezone handling\n- Plotting improvements\n\n### Version 0.7 (2020.09.05)\n\n- Built-in json serialization\n- Added \"flat\" growth option\n- Bugfixes related to `holidays` and `pandas`\n- Plotting improvements\n- Improvements in cross validation, such as parallelization and directly specifying cutoffs\n\n### Version 0.6 (2020.03.03)\n\n- Fix bugs related to upstream changes in `holidays` and `pandas` packages.\n- Compile model during first use, not during install (to comply with CRAN policy)\n- `cmdstanpy` backend now available in Python\n- Python 2 no longer supported\n\n### Version 0.5 (2019.05.14)\n\n- Conditional seasonalities\n- Improved cross validation estimates\n- Plotly plot in Python\n- Bugfixes\n\n### Version 0.4 (2018.12.18)\n\n- Added holidays functionality\n- Bugfixes\n\n### Version 0.3 (2018.06.01)\n\n- Multiplicative seasonality\n- Cross validation error metrics and visualizations\n- Parameter to set range of potential changepoints\n- Unified Stan model for both trend types\n- Improved future trend uncertainty for sub-daily data\n- Bugfixes\n\n### Version 0.2.1 (2017.11.08)\n\n- Bugfixes\n\n### Version 0.2 (2017.09.02)\n\n- Forecasting with sub-daily data\n- Daily seasonality, and custom seasonalities\n- Extra regressors\n- Access to posterior predictive samples\n- Cross-validation function\n- Saturating minimums\n- Bugfixes\n\n### Version 0.1.1 (2017.04.17)\n\n- Bugfixes\n- New options for detecting yearly and weekly seasonality (now the default)\n\n### Version 0.1 (2017.02.23)\n\n- Initial release\n\n## License\n\nProphet is licensed under the [MIT license](LICENSE).\n","Prophet 是一个用于时间序列数据分析的工具，特别适用于具有多重季节性特征且增长趋势为线性或非线性的数据。其核心功能包括自动识别并拟合年、周和日等不同周期的季节性模式以及节假日效应，同时对缺失值和趋势突变有较好的鲁棒性。Prophet 采用 Python 和 R 语言实现，并遵循 MIT 许可证开源发布，适合需要进行高质量预测分析的场景，如销售预测、网站流量预测等领域。",2,"2026-06-11 02:50:04","top_language"]