Transparency in data science

ArcGIS Notebooks is Esri’s answer for streamlining complex geospatial analysis. It is based on Python and Project Jupyter. Use ArcPy, ArcGIS and hundreds of open source Python libraries to support any geospatial research.

Data science, like deep learning allows us to discover meaning and trends in complex data and spatial systems. Examples are pattern analysis of aerial imagery or seeking trends in terabytes of network related IoT-data. Results from such research can be great. However, too often the underlying scientific algorithms are presented as a black box. We are asked to simply ‘trust the system’.  

Detecting specific tree species using ArcGIS.
Source: Geopunt.be

Earth scientists are trained to use GIS and scripting to analyse multi-variate and high-volume spatial data and services. And indeed the methods and results can be complex. The need for communication is therefore always present, as is the expectancy to justify each decision made. ArcGIS Notebooks is designed to facilitate and manage this process.

With the ergonomic interface you gradually and visually elaborate your study. Over 300 Python libraries are at your disposal, as are ArcPy and Esri’s advanced image processing and geo-analytics toolkits*.

The ArcGIS Notebooks interface. Source Esri.

On top of that, ArcGIS Notebooks offers the possibility to insert intermediate results. You can use maps, graphs, and tables while constructing your Python script. By showing intermediate results it will be quite straightforward to verify new models. In addition, when presented your research will prove to be fully explainable.

Thanks to ArcGIS Notebooks the transparency of spatial data science remains guaranteed, even in later stages when calculations are getting multi-dimensional and often rather complex.

ArcGIS Notebooks has been available for ArcGIS Desktop and ArcGIS Enterprise for over a year. Since recently it is also available via ArcGIS Online. Indeed, you can start your data science project based on an affordable subscription.

Do contact us if you want more information about ArcGIS Notebooks or when you want to discuss a specific study.

So, start building your deep learning algorithms today, or have it developed and expect clear reporting about the scientific methods used.

* Availability of specific tools may require additional licensing. Do contact us for more information.