# DataScienceInteractivePython: Interactive Notebook Dashboards for Learning Data Science > GeostatsGuy's DataScienceInteractivePython is a collection of Jupyter notebooks that turn statistics, machine learning and geostatistics concepts into ipywidgets dashboards. It is a teaching resource, not a library, and that distinction shapes everything about how you use it. **GeostatsGuy/DataScienceInteractivePython** — Python interactive dashboards for learning data science - Repository: https://github.com/GeostatsGuy/DataScienceInteractivePython - Stars: 2,584 · Forks: 465 - Language: Jupyter Notebook - License: CC-BY-SA-4.0 - Published: 2026-09-28 · Updated: 2026-09-28 · Language: en - Canonical page: https://hysenlabs.com/projects/geostatsguy-datascienceinteractivepython ## What DataScienceInteractivePython Is Actually For This is not a library you import. It is a set of 60-plus Jupyter notebooks, one per concept, each built as an ipywidgets dashboard so a learner can drag a slider and watch a statistic, a model fit or a spatial realization change. The README states the motivation plainly: the author, Michael Pyrcz at the University of Texas at Austin, writes a new dashboard when students struggle with a concept, so they can learn by playing with the statistics, models or theoretical concepts. The audience is therefore narrow and specific. It is students in data analytics, geostatistics and machine learning courses, plus working engineers and geoscientists who want the intuition behind variograms, declustering, bootstrap or principal component analysis without reading a derivation first. The repository summary lists the coverage: Bayesian and frequentist statistics, confidence intervals and hypothesis testing, Monte Carlo and bootstrap methods, cluster analysis, hyperparameter tuning and overfit models, spatial data debiasing, variogram calculation and modeling, spatial estimation and simulation, and decision making under uncertainty. If you arrived looking for a pip-installable package that adds interactive plots to your own application, this is the wrong repository. The notebooks are the product. The interactive widgets exist to explain a concept to a human, not to be embedded in a pipeline. ## How the Notebooks and Widgets Fit Together Each notebook is self-contained. It imports the scientific stack (NumPy, Pandas, MatPlotLib, SciPy), imports ipywidgets for the controls, and in the geostatistics notebooks imports GeostatsPy for the spatial algorithms and functions. The datasets are not stored in this repository; the README says the required datasets live in the separate GeoDataSets repository and are linked in the workflows. That means every notebook that reads data depends on a network fetch from another GitHub repository at run time. The interaction model is the standard ipywidgets one: a widget callback recomputes a result and redraws a MatPlotLib figure in place. Because the computation happens in the kernel, the responsiveness of a dashboard is bounded by how fast the underlying algorithm runs in Python. The README gives one concrete reason for the pinned Python version: Python 3.7.10 is required due to the dependency of GeostatsPy on the Numba package for code acceleration. Numba compiles selected functions, and its supported Python range is narrower than the language's, which is why an educational repository ends up specifying an exact interpreter version rather than a minimum. The file names are the map of the content. Interactive_Central_Limit_Theorem.ipynb, Interactive_Bootstrap.ipynb, Interactive_Declustering.ipynb, Interactive_PCA.ipynb, Interactive_Simple_Kriging.ipynb and Interactive_Overfit.ipynb sit side by side with narrower ones such as Interactive_QQ_Plot.ipynb, Interactive_Spurious_Correlations.ipynb and Interactive_String_Effect.ipynb. There is no index file, no package metadata beyond the release tags, and no test suite. You find the notebook you need by reading the directory listing. ## Running a First Dashboard from the Repository The README offers two paths. The fast one is Binder: the badge links to a mybinder.org container built from the repository, which the author added because some students had trouble setting up local environments. Clicking it launches the notebooks in a browser with no local install. The slow path is a local environment containing the packages the README lists. For the local route, the README names a minimum environment rather than a sequence of commands. It specifies Python 3.7.10 and the packages MatPlotLib, NumPy, Pandas, SciPy, ipywidgets and GeostatsPy, with GeostatsPy available from PyPI. Because the README does not give install commands, the exact tooling is your choice; what matters is that the interpreter version and the package set match that list. Once that environment exists, open a notebook from the repository root. The central limit theorem dashboard is a reasonable first stop because it needs no external dataset. The README does not publish a launch command, so use whatever interface you normally start Jupyter with, and open the file by name. What you should see is a notebook with widget controls above a MatPlotLib figure. Moving a control re-runs the cell's callback and redraws the plot. Notebooks that read data will instead fail at the data-loading cell until the linked GeoDataSets files are reachable, which is the first thing to check if a dashboard opens but stays blank. ## Where the Repository Gets in Your Way The Python 3.7.10 pin is the sharpest constraint. That interpreter reached end of life in 2023, and the README does not describe a migration path or a newer supported combination. On a current machine you will either build the old environment, which means an older conda or a container, or you will install the packages on a modern interpreter and accept that GeostatsPy may not import cleanly because of Numba. The README does not document what happens in that second case, so treat it as untested ground rather than a supported configuration. The external dataset dependency is the second friction point. Because the data lives in GeoDataSets and is linked from the notebooks, a change or a rename in that repository can break a lesson without anything changing here. The README does not document a fallback copy of the data or a pinned commit. There is also no packaging. No setup.py, no pyproject.toml, no importable module. If your goal is a reusable interactive component for a product, a dashboard framework such as Panel or Dash is the right class of tool, because those are libraries you build against rather than notebooks you open. This repository is the opposite: content you read and manipulate. Finally, the release history is thin. Two tags exist, DataScienceInteractivePython 0.0.1 from 2024-07-05 and v0.2 from 2025-04-07, and the README still carries 0.0.1 in its title and citation. The last push to the repository was on 2026-05-15, so work has continued, but the versioning does not track it in a way you can depend on. ## How It Compares with a General Teaching Stack The obvious alternative is a general-purpose interactive Python stack: Jupyter widgets plus a plotting library, assembled yourself around whatever textbook or dataset you already use. The difference is in what has been pre-decided. Here, the concept, the widget controls, the plotting code and the pedagogical sequencing are all already written for each topic. Building the same central limit theorem or bootstrap demo from scratch is an afternoon of work that teaches you plotting more than it teaches you the theorem. A second alternative, for the spatial material specifically, is GeostatsPy on its own. GeostatsPy is the library this repository depends on; it provides the variogram, kriging and simulation functions. If you want to compute a variogram in your own script, install GeostatsPy and call it. What you lose is the widget layer that makes the parameters visible. The notebooks are valuable precisely because they wrap those functions in controls, which is a different deliverable from the functions themselves. For pure statistics education, a notebook-based course repository is the closer comparison. Those tend to favor prose and exercises over live manipulation. This repository inverts that: the interaction is the explanation, and the surrounding text is thin. If your students learn better from reading a derivation than from dragging a slider, the format will feel like overhead. ## Licence, Maintenance and the Cost of Upgrading The repository is licensed CC-BY-SA-4.0. That is a content licence, not a software licence, and it carries a share-alike condition: adaptations distributed to others must be released under the same terms, with attribution. The README provides the citation to use, Pyrcz, Michael J. (2021), DataScienceInteractivePython, Zenodo, with the DOI 10.5281/zenodo.5564966. For a course, reusing a notebook as-is with attribution is straightforward. Forking the notebooks, editing them for your own curriculum and publishing the result is the case where the share-alike term matters, and it is worth reading the licence text rather than assuming. This is a description of the terms, not legal advice. On maintenance, the facts are limited. The repository is not archived, and the last push was on 2026-05-15. Two releases exist, the most recent being v0.2 on 2025-04-07. Beyond that, the README documents no support policy, no deprecation schedule and no compatibility matrix past the Python 3.7.10 requirement. The upgrade cost is therefore not a version bump you plan for; it is the cost of rebuilding an environment when the pinned interpreter and the Numba-dependent GeostatsPy combination stops being installable through normal channels. Budget for a container image of the working environment if you intend to teach from these notebooks across several semesters, because that is the only way to freeze the combination the README describes. ## Conclusion Adopt this repository if you teach or study data analytics, geostatistics or machine learning and want students to manipulate distributions, variograms and models instead of watching a lecture. Skip it if you need a maintained Python package with a stable API, tests and semantic versioning; the notebooks are the deliverable and the README lists no test suite or release cadence beyond the v0.2 tag from 2025-04-07. Before committing to it in a course, open the notebook that matches your topic, confirm the GeoDataSets datasets it links to still resolve, and check that your Python version matches the 3.7.10 the README specifies for GeostatsPy and Numba. ## FAQ ### What is DataScienceInteractivePython? It is a repository of Python interactive dashboards, delivered as Jupyter notebooks, for learning data science. The README describes it as a set of dashboards the author builds when students struggle with a concept, covering statistics, machine learning and geostatistics topics. ### How can I create interactive dashboards in Python like these? The notebooks use ipywidgets for plot interactivity alongside MatPlotLib for plotting, with widget callbacks recomputing results and redrawing figures in the kernel. The README lists ipywidgets as part of the minimum environment. ### What Python version does DataScienceInteractivePython require? The README specifies Python 3.7.10, stating this is due to the dependency of GeostatsPy on the Numba package for code acceleration. It also lists MatPlotLib, NumPy, Pandas, SciPy, ipywidgets and GeostatsPy as the minimum environment. ### Where do the datasets used by DataScienceInteractivePython come from? The README states that the required datasets are available in the GeoDataSets repository and are linked in the workflows. They are not stored in this repository, so the notebooks fetch them from there. ### Can I run DataScienceInteractivePython without installing anything locally? Yes. The README provides a Binder badge that launches a container from the repository on mybinder.org, which the author added because some students had issues setting up local computing environments. ## Sources - [GeostatsGuy/DataScienceInteractivePython on GitHub](https://github.com/GeostatsGuy/DataScienceInteractivePython) - [Issues](https://github.com/GeostatsGuy/DataScienceInteractivePython/issues) - [License: CC-BY-SA-4.0](https://github.com/GeostatsGuy/DataScienceInteractivePython/blob/main/LICENSE) - [README](https://github.com/GeostatsGuy/DataScienceInteractivePython/blob/main/README.md) - [Releases](https://github.com/GeostatsGuy/DataScienceInteractivePython/releases) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/geostatsguy-datascienceinteractivepython