Open Source Infrastructure Risk Analytics (osira)¶
Welcome to the documentation for osira!
Description¶
These docs provide an overview of the Open Source Infrastructure Risk Analytics (osira) codebase written in Python.
The aim is to be able to quantify the direct and indirect impacts of infrastructure cascading failure, such as from cyber-attacks on electricity assets.
Citation¶
Oughton, E. J. et al. (2019) Stochastic Counterfactual Risk Analysis for the Vulnerability Assessment of Cyber-Physical Attacks on Electricity Distribution Infrastructure Networks, Risk Analysis, 39(9), pp. 2012–2031. https://doi.org/10.1111/risa.13291.
Kelly, S. et al. (2016) Integrated Infrastructure: Cyber Resiliency in Society, Mapping the Consequences of an Interconnected Digital Economy. Cambridge: Cambridge Centre for Risk Studies.
Statement of Need¶
Disruption in electricity supply has major ramifications for both society and the economy. Risk analysts working in the insurance sector have a major interest in trying to understand the potential business interuption impacts.
Indeed, catastrophic events such as cyber-attacks are both a major risk management issue and a huge business opportunity for different types of insurers.
However, it is surprising that we lack open-source models to help quantify these risks, providing strong motivation for the content of the osira repository.
Setup and Configuration¶
All code for itmlogic is written in Python (Python>=3.7).
See requirements.txt for a full list of dependencies.
Installing via conda¶
The recommended installation method is to use conda to handle packages and virtual environments. The conda-forge channel also has a host of pre-built libraries and packages.
Create a conda environment called osira:
conda create –name osira python=3.7 gdal
Activate it (run this each time you switch projects):
conda activate osira
First, you need to install necessary packages, which at a minimum, is geopandas:
conda install geopandas pytest matplotlib seaborn
Then clone this repository and run:
python setup.py install
Or if you want to develop the package:
python setup.py develop
And, also should you want to run the tests:
python -m pytest
Quick Start¶
To quickly get started using synthetic data run this:
python scripts/demo.py
Followed by using the vis.py script to visualize the results:
python vis/vis.py
Background and funding¶
The approach has been developed over many years at numerous institutions:
2015-2017: Cambridge Centre for Risk Studies, University of Cambridge
2017-2020: Environmental Change Institute, University of Oxford
2020-2021: Geography and Geoinformation Sciences, George Mason University
We would like to thank UKRI, specifically the Engineering and Physical Sciences Research Council for support via grant EP/N017064/1.