Predicting a Crises

Predicting a Crises

Assignment for Information Management and Organizational Change, MSc Business Information System, University of Amsterdam

4796c2853f14f3c6f0825070f1d2e052?s=128

Adriano Martins

May 22, 2012
Tweet

Transcript

  1. CRISIS predicting a Adriano Martins Bjorn Burscher Hamid Nasiri Nikolas

    Kourtis Peter Bouwdewijn
  2. predicting a CRISIS Crisis Data Analysis Examples 2

  3. CRISIS "A significant business disruption that stimulates extensive news media

    coverage. The resulting public scrutiny will affect the organization’s normal operations and also could have a political, legal, financial and governmental impact on its business." 3 Institute for Crises Management
  4. CRISIS Sudden Impossible to predict Variables out of scope Too

    complex Smoldering Possible to predict Variables in scope Complicated 4
  5. CRISIS sudden 39% smoldering 61% management 50% other 18% employee

    32% source: Institute for Crisis Management 5 Sudden vs. Smoldering Origins (smoldering)
  6. smoldering crisis classification CRISIS 1. ❖ internal business problem ❖

    can be resolved by a management assigned to it 2. ❖ internal business problem ❖ can be resolved by the management assigned to it, with the help from higher management 3. ❖ internal business problem ❖ potential of going public 4. ❖ serious situation ❖ likely to spread to the public ❖ direct and strong impact on the business 6
  7. CRISIS “Wal-Mart Meeting Videos Were a Smoldering Crisis That Could

    Have Been Prevented” how could it be predicted? Wal-Mart finishes a long business relation with Flager Flager’s looses 95% of it’s income In order to survive, Flagler became a video archive for anyone looking for evidence in suits against the Wal- Mart Normaly, Wal-Mart was constantly sued Flagler charges 200€/h for video research There was no contract, so Wal-Mart had no rights on the tapes.
  8. The Company Competitors Technological Customers Law etc data CRISIS The

    World Governments Conflicts Markets Weather Technology Disasters etc 8 ta
  9. crises and socio-economic systems CRISIS Organizational crises result from complex

    dynamics within social economic systems Using data analysis, we might understand these dynamics (partly) Analyzing data can help avoiding & managing... ❖ socio-economic crises, ❖ systematic instabilities, ❖ and other contagious cascade-spreading processes 9
  10. complexity in social economic systems CRISIS more complex social economic

    systems due to... ❖ non-linear dynamics (small events can lead to massively consequential results) ❖ increasing interdependence (globalization, development of global communication networks, etc) 10
  11. CRISIS 11 chaos theory Butter-fly effect highly sensible to input

    variables Link: http:// www.youtube.com/ watch?v=QCsLTPGpEfw
  12. chaos implications CRISIS “Predicting that a storm is coming, but

    not being able to predict it will form” (Mandelbrot) ❖ Impossible to know your complete environment and interrelated variables “One alteration destroys years of conformation” (Taleb) ❖ Knowledge about your environment can be swept away by one alteration 12
  13. Data Increase in (decentralized) data sources Digital traces Increasing rates

    of data generation Increasing interconnectedness and data exchange 13 increasing amounts of it
  14. Data Data explosion problem due to developments in ... ❖

    automated data collection tools ❖ database technology Mature developments in database technology ➡ Tremendous amounts of data stored in databases, data warehouses and other information repositories ➡ Need for automatic knowledge extraction Solution: Data Mining 14 we are drowning in data, but starving for knowledge! “
  15. Data〜~Mining ❖ Extraction of interesting knowledge (rules, regularities, patterns, constraints)

    from data in large databases Goal ❖ Determining relationships among "internal" factors and "external" factors in order to 15 what is it?
  16. Data〜~Mining A massive mining of socio-economic data can... ❖ reduce

    gaps in our knowledge and understanding of techno-social- economic-environmental systems ❖ predict crises or identify systematic weaknesses, and help to avoid or mitigate impacts of crises ❖ enable real-time sensing and data collection (¨reality mining¨) to reduce mistakes and delays in decision making 16 motivation
  17. potential applications Data〜~Mining Market Analysis, Target Marketing, CRM ❖ Resource

    planning ❖ Predicting socio-economical developments ❖ stock market & resource market ❖ environmental change ❖ cultural, technological and political developments 17
  18. from data to knowledge Data〜~Mining Data ❖ Internal (operational &

    transactional data, customer data) vs external (web, news media, social media, metadata) Information: ❖ Pattern, associations, relationships among data can provide information Knowledge: ❖ Information can be converted into knowledge about historical patterns and future trends 18
  19. steps of a data mining process Data〜~Mining 19

  20. real-time knowledge mining Data〜~Mining Until now ❖ gathering data ⇢

    analyzing data But ❖ contemporary challenges require real-time reactions to changes in respective environment Continuous data analysis over streaming data 20
  21. the wisdom of crowds Data〜~Mining Real time prediction of user

    activity Search results, social network posts, tweets, blogs, massive multiplayer gaming Trend prediction, identification of opinion leaders, trend setters & innovators predicting election outcomes, economic developments, epidemies 21
  22. 22 Data〜~Mining The Prediction API enables you to make your

    smart apps even smarter. The API accesses Google's machine learning algorithms to analyze your historic data and predict likely future outcomes. Using the Google Prediction API, you can build the following intelligence into your applications. Link http://www.youtube.com/watch?v=u39rCNFWDEA google flu trends
  23. data-analysis Data〜~Mining Descriptive Analysis ❖ finding distinctive features in existing

    data Predictive Analysis ❖ deriving trends from data ❖ interest in temporal patterns, rules, dependencies, regularities Statistical Methods ❖ Clustering ❖ Classification ❖ Regression ❖ Neural Networks ❖ Machine Learning ❖ etc. 23
  24. CRISIS 24 predicting of sudden crisis no certainty! link http://www.youtube.com/

    watch?v=SIFj_YMhokg
  25. CRISIS 25 lesson to the professional Political influence Link http://www.youtube.com/watch?v=SIFj_YMhokg

  26. google prediction api cloud based machine learning tool can do

    ✓sentiment analysis ✓purchase predictions Link http://www.youtube.com/watch? v=u39rCNFWDEA CRISIS 26
  27. CRISIS predicting a Adriano Martins Bjorn Burscher Hamid Nasiri Nikolas

    Kourtis Peter Bouwdewijn