GIS Meetup & Educational Seminar Summary #2
Presented on: Nov 2nd, 2019
Presenters:
- Shannon Kalisky, Product Manager, Analytics and Data Science, Esri
- Atma Mani, Lead Product Engineer, ArcGIS Python API, Esri
- Lipika Gimmler, Product Marketing Specialist, Spatial Analysis and Data Science, Esri
External link:
https://www.geotechcenter.org/webinar-archives.html
Overview
Geospatial data science is a subset of data science which covers a spectrum of activities including data engineering, visualization and exploration, spatial analysis, machine learning and artificial intelligence, big data analytics, modeling and scripting, and sharing and collaboration.
Geo-enrichment
Geo-enrichment is a process which connects data with location. Adding spatial analysis to data helps us to understand the moment, see the big picture, coordinate and connect, and plan for disasters. However, maps are subjective and can be made to tell many different stories.
Jupyter Notebooks
Jupyter notebooks are a widely used tool in the field of data science. Notebooks are used to create reproducible research and increase transparency in analytical processes. Notebooks are also extremely helpful for data engineering and exploration.
ArcGIS Notebooks
ArcGIS Notebooks bring Jupyter notebooks to ArcGIS. ArcGIS Notebooks are Jupyter notebooks hosted in the cloud which enable users to centrally manage packages and libraries and share their notebooks via ArcGIS Online.
ArcGIS Notebooks are available in the ArcGIS Pro, Enterprise, and Online products.
ArcGIS Notebooks come with a gallery of pre-built notebooks which provide templates so that you do not have to start from scratch. Notebooks help users engineer data for analysis and machine learning, perform exploratory analyses, conduct spatial analyses and data science, add context to build stories, and increase the reproducibility and transparency of workflows.
Notebook Applications
Notebooks can be used for data cleaning/wrangling, exploratory analysis, building models, and automating tasks. Some of the capabilities of ArcGIS Notebooks include the ability to share notebooks as items, import existing .ipynb files into ArcGIS and used them with ArcGIS notebooks, automatically insert snippets of Python code into analyses, and use open source Python libraries via ArcGIS API for Python and ArcPy. Using the profiler tool via the ArcGIS API for Python, spatial data can be added to Pandas data frames as a new SHAPE column in the data frame.
Demonstration
A demonstration of data engineering using ArcGIS Notebooks was given. In the demonstration 169 years of hurricane data was cleaned up to create 12,000 hurricane tracks which were then loaded into ArcGIS as polylines. A second demonstration was given of density-based clustering of Denver food store data using ArcPy.
Reaction
I am excited by the power that comes from combining spatial data and data science tools such as Jupyter notebooks. I am already familiar with Jupyter notebooks, so it was great to see how this technology integrates with ArcGIS. I would like to work through some of the sample notebooks in the gallery when I have time. In addition, I am interested in taking the MOOC “Spatial Data Science – The New Frontier in Analytics” mentioned in the presentation. Although this MOOC is not currently available, I have registered my interest with Esri.