Storytelling with Data: Data Exploration and Visualization
NSF Unidata hosted the Unidata Users Workshop, Storytelling with Earth System Science Data: Challenges and Opportunities for Effective, Ethical, and Reproducible Science, on 5-8 June 2023 in Boulder, Colorado. This selection of presentations focuses on data exploration and visualization.
Presentations
Exploratory Data Analysis: A Conceptual Overview
Drew Camron Exploratory Data Analysis (EDA) helps answer foundational questions about a dataset before addressing specific scientific problems. It involves understanding how data are organized, assessing tool readiness, verifying data provenance, and evaluating dataset suitability. EDA is an iterative process applied throughout the scientific workflow. This session introduces key EDA concepts and presents example scenarios to prepare participants for future applications. |
|
Exploratory Data Analysis: Tabular Data
Thomas Martin This session covers basic Python syntax using the pandas package to visualize and describe tabular datasets. |
|
Exploratory Data Analysis: Multidimensional Data
Jon Thielen This session demonstrates how to perform exploratory data analysis on multidimensional datasets, such as netCDF, using Jupyter Notebooks. The xarray Python package is used to examine metadata, calculate statistics, and preview array-based data. |
|
How Do I Find a Dataset of Interest and Access It?
Douglas Schuster This session reviews strategies for discovering datasets related to ocean and atmospheric science and explores the various platforms that host these datasets. Participants examine access methods, ranging from standard data transfers to interactive visualization and analysis tools. |
|
Working with Data from Multiple Domains
Paul Fremeau This session explores how forensic meteorologists address challenges in real-world accident investigations by integrating data from multiple domains. Examples include flight data recorders (FDR), Aviation Weather Center forecasts, HRRR and WRF model data, satellite and radar data, pilot reports, and eyewitness accounts. Participants examine the tools used in these investigations and how to combine them to create compelling multidisciplinary stories. |
|
The Intersection of Earth Systems Sciences and Social Science Data
Andrea Schumacher This session examines challenges in interdisciplinary research, particularly in integrating data from Earth System Science and Social Science. Differences in temporal, spatial, and formatting requirements across disciplines can hinder meaningful integration. The presenter shares personal experiences related to hurricane risk perception and response, and participants explore strategies to address and overcome common integration obstacles. |
|
Rebuild the National Weather Service
Kim Klockow-McClain In this session, participants take on a scenario-based challenge: imagine a world without existing NWS Weather Forecast Offices (WFOs) and propose a data-informed plan to determine where new ones should be located. Participants use both meteorological and societal datasets to support their reasoning, considering factors like service equity, hazard climatology, and impact distribution. The session deepens understanding of the systemic challenges WFOs face and demonstrates how multiple datasets can be combined to address institutional inequities. |
|
Meteorological and Societal Data: A Case Study of an EF4 Tornado Event
Stephen Strader This session presents a case study of an EF4 tornado event using data from atmospheric, social, and geospatial sciences. Participants learn how to identify and analyze multiple data sources, use plotting methods, and interpret results within a geospatial context. The session emphasizes how integrated datasets provide insights into high-impact weather hazards. |
|
The Unreasonable Effectiveness and Frustration of Metadata in Data-Intensive Analysis
Deepak Cherian As the volume and variety of climate datasets continue to grow, insight often lies at their intersection. This session focuses on the cf-xarray Python package and demonstrates how metadata can streamline complex analysis pipelines. It also highlights common challenges that arise when relying on metadata for scientific workflows. |
|
Data Visualization in the Private Sector
Nick Lilja This session explores how the private sector collects meteorological data, creates visualizations, and communicates forecasts to its audiences. Emphasis is placed on simplicity, clarity, and specificity. The session illustrates how visualizations serve as powerful tools for condensing complex data into accessible insights. After all, one picture can be worth a thousand words. |
|
Details
Learning Series
Storytelling with Data
Intended Audience
Earth Systems scientists and students
Format
Video, Jupyter Notebook