Heterra

Heterra

Advanced spatial analysis. One lightweight GIS app.

前沿空间分析方法,一款轻量级 GIS 软件

Heterra is free, lightweight software for exploring spatial heterogeneity, uncovering patterns, building predictions, and evaluating models—with interactive maps, charts, and guided tutorials for every method.

Heterra 是一款免费、轻量的空间分析软件,用于探索空间异质性、发现空间模式、构建预测模型并评估模型表现。每种方法均配有交互式地图、统计图和引导式教程,并可导出适用于学术论文与研究报告的高质量图件。

Download for Mac下载 Mac 版

Apple silicon · macOS 13 or later · English and 中文

Apple 芯片 · macOS 13 或更高 · English 和中文

Heterra's main window showing a map of NDVI change in seven natural-break classes

A native Mac app

原生 Mac 应用

Open shapefiles, GeoTIFF, GeoPackage, GeoJSON, CSV and File Geodatabase. Style and explore layers, then run a method in the same window, with no R or Python to install.

可打开 Shapefile、GeoTIFF、GeoPackage、GeoJSON、CSV 和 File Geodatabase。在同一个窗口里设置符号、查看图层、运行方法,不用安装 R 或 Python。

Results checked against R

结果与 R 核对

Each method is tested against its R reference implementation. Every tutorial compares its results with the reference on your own Mac, and every run can be exported as an R script.

每个方法都与其 R 参考实现对照测试。教程会在你自己的 Mac 上把结果与参考结果比较,每次运行都可以导出为 R 脚本。

Figures ready for papers

可直接用于论文的图

Export charts and page layouts as vector PDF and SVG, or as PNG at print resolution, with colours from RColorBrewer. Keep the data behind each chart as CSV.

统计图和版面可导出为矢量 PDF、SVG,或印刷分辨率的 PNG,颜色来自 RColorBrewer。每张图背后的数据也可以导出为 CSV。

  • Set up in steps, run where it counts

    分步设置,需要时才运行

    Each tool walks through its settings one step at a time. Run appears only at the steps that compute something.

    每个工具按步骤设置,只有需要计算的步骤才有运行按钮。

  • A tutorial in every tool

    每个工具都有教程

    Load an example, follow the steps and read the results, with a guide in English and Chinese.

    载入案例,跟着步骤操作并解读结果,手册有中英文。

  • Charts linked to the map

    统计图与地图联动

    Where a chart lists a layer's categories, click one to select its features on the map.

    统计图列出图层的类别时,点击一个类别就能在地图上选中对应要素。

Heterra's Chart page showing the degree of spatial interpretability of five models

Methods方法

Spatial heterogeneity空间异质性

OPGD
Optimal parameters-based geographical detector最优参数地理探测器 Finds the discretization of each continuous factor that best explains y, then runs the factor, interaction, risk and ecological detectors. 为每个连续因子寻找最能解释 y 的离散化方案,再运行因子、交互、风险和生态探测。
Song et al. 2020
Wang et al. 2010
LISP
Local indicator of stratified power局部分层影响 The geographical-detector q inside a window around every location, for each variable, variable pairs and all variables together. 在每个位置周围的窗口内计算地理探测器 q,分别对单个变量、变量对和全部变量计算。
Hu et al. 2024

Spatial pattern空间模式

GC
Geocomplexity地理复杂性 How strongly each location departs from the smooth spatial dependence of its neighbourhood, per variable. 逐变量衡量每个位置偏离其邻域平滑空间依赖的程度。
Zhang et al. 2023
SDA
Second dimension of spatial association第二维度空间关系 Quantiles of each covariate around every sample become new variables; the stable ones are selected and their gain cross-validated. 把每个样本周围协变量的分位数变成新变量,选出稳定的变量,并用交叉验证说明精度提高多少。
Song 2022
SDO
Second-dimension outliers第二维度异常 Local outlier strengths of each covariate around every sample become new variables; the stable ones are selected and their gain cross-validated. 把每个样本周围协变量的局部异常强度变成新变量,选出稳定的变量,并用交叉验证说明精度提高多少。
Ren et al. 2026a
Ren et al. 2026b
Song 2022

Spatial prediction空间预测

GCF
Generalized covariate field广义协变量场 Expands covariates into spatial-pattern and neighbourhood-distribution features, selects stable predictors and predicts. 把协变量扩展为空间模式与邻域分布特征,选出稳定的预测变量并预测。
Song 2026
GOS
Geographically optimal similarity地理最优相似度 Predicts a variable at new locations from the samples whose covariate configuration is most similar, with a prediction uncertainty. 用协变量配置最相似的样本预测新位置上的变量,并给出预测不确定性。
Song 2022

Model evaluation模型评估

DSI
Degree of spatial interpretability空间解释度 How much of the spatial autocorrelation and spatial heterogeneity in y each model's residuals no longer carry. 衡量每个模型的残差中不再含有 y 的多少空间自相关与空间异质性。
Liu et al. 2026
SIE
Spatial information explained空间信息解释率 How much of the spatial information in y a prediction model explains, from out-of-fold predictions. 根据交叉验证的样本外预测,衡量模型解释了 y 中多少空间信息。
Song 2026

Data tools数据工具

CSV to points · Reproject · Raster to grid table · Join · Field calculator · Clip

CSV 转点 · 重投影 · 栅格转格网表 · 连接 · 字段计算 · 裁剪

Download下载

Coming soon即将发布

The first public release is being prepared.

第一个公开版本正在准备中。

  • Requirements. A Mac with Apple silicon and macOS 13 or later.
  • 系统要求。Apple 芯片的 Mac,macOS 13 或更高。
  • Install. Open the .dmg and drag Heterra to Applications.
  • 安装。打开 .dmg,把 Heterra 拖到"应用程序"。
  • First launch. Heterra is not notarised by Apple, so macOS blocks it the first time. Open System Settings → Privacy & Security and choose Open Anyway.
  • 第一次打开。Heterra 没有经过 Apple 公证,macOS 第一次会拦截。请打开"系统设置 → 隐私与安全性",选择"仍要打开"。

Cite引用

Song, Y. (2026) Heterra [software]. https://heterra.github.io

Please also cite the paper of each method you use, linked above. Heterra shows the full reference in every tool's panel.

也请引用你所用方法的论文(链接见上)。每个工具的面板里都有完整的文献信息。

Feedback反馈

If you have any suggestions, comments or questions, please email yongzesong.heterra@gmail.com.

如有任何建议、意见或问题,请发邮件到 yongzesong.heterra@gmail.com。