Date post: | 27-Jan-2015 |
Category: |
Economy & Finance |
Upload: | kimmo-soramaki |
View: | 106 times |
Download: | 0 times |
Financial Network Analytics
Kimmo Soramä[email protected]
Presentation at Bank of Canada
Workshop on Understanding Financial Institutions and Markets: Theory and Data
2 June 2011
Perspective
Understanding financial institutions and markets
Theory Data
Focus on the tools and processfor research and policy
Katsushika Hokusai. The great wave off Kanagawa ~1830
Eratosthenes' map of the known world, c. 194 BC
Go is a board game that originated in China more than 2,000 years ago.
Data deluge
• Data tsunami - digital information is doubling every 1.2 years
• Regulatory response to recent financial crisis was to strengthen macro-prudential supervision with mandates for more regulatory data
• The challenge will be to understand and analyze the data (public, commercial and regulatory)
• Data enables “Analytics based policy”, i.e. the application of computer technology, operational research, and statistics to solve problems
Network maps
• Recent financial crisis brought to light the need to look at links between financial institutions
• Way to present data. Mapping of the financial system has only begun
• Reduce uncertainty about interconnections
Soramaki, Bech et al (2006)
Garrett, Mahadeva and Svirydenska (2011)
Intelligence
• Financial crisis are different and rare
• Pattern recognition is hard for computers -> the best Go programs only manage to reach an intermediate amateur level
• The patterns to be recognized must be frequent enough for computers to learn
• A solution is to augment human intelligence (in contrast to AI)
Roots of the work• Bof-PSS2
– Bank of Finland, 1997-– Payment system simulator used in ~60 central
banks
• Loki– Sandia National Laboratories, 2004-– Toolkit for network analysis and ABM
• Sponsored by Norges Bank, collaborative efforts with other central banks
Method and features
Data validation
Analysis and modeling Visualization
Data tools in FNA
ValidateCheck field content against
expected values or rules for values
CrosscheckEnsure that different datasets are
consistent with each other
Explore and extractExplore data by enumerating
field values and extract subsets of data. Operate directly on
compressed files.
AnonymizeAnonymize fields consistently
across datasets
Data validation
Network analysis in FNA
AnalyzeMost common network metrics for
centrality, components and connectivity. Aggregate properties.
GenerateGenerate random networks for
comparing results
OperateDelete, add and merge vertices and
networks. Transpose and symmetrize networks.
VisualizeInteractive graph layouts,
geographic layouts + all standard charts
Network analysis
Network visualization in FNA
PlotAll basic chart types (pie, line, bar, scatter), Interactive graph layouts,
geographic layouts
BrowseBrowse networks visually,
visualizations are auto-generated for all networks
InteractShow metadata, interact, move and
highlight, zoom
Share and extendShare independent html
documents, publish & find on web
Visualization
Demo
Versions
– Web – Cloud service– REST API (for accessing FNA engine from web applications)
– Intranet– Operative use– Integration to databases and user management– Browser based GUI
– Desktop – By demand, academic use– Command line version available– Integration with external programs like Matlab, Stata, Excel, etc.
Availability
• Private beta in July 2011
• Official release in August 2011
• Core (graph engine, algorithms, visualizations) is open source (GPL)
• Desktop version is free for academic use
Vision
• Develop tool for explorative analysis of the financial system
• Provide a platform for building agent based and simulation models
• Make advances in research available to policy
• Integrate data feeds and available databases