Downloaded 11 times
Uploaded byMarten den Haring
Analytics Summit 2013
The document discusses the integration of analytics in organizations, emphasizing the importance of leveraging various data tools and technologies to extract insights and create data products. It highlights the challenges of compliance in the heavily regulated financial industry, illustrated through examples of significant trading losses due to unauthorized activities. The text underlines the utility of human language as a rich data source for understanding trader behavior and improving risk management through the development of key risk indicators.
Analytics Summit 2013
- 1. A Data Cocktail Dirty,with a Compliance Twist
- 2. How to effectivelyleverage analytics in your organization? Stock your bar with data & tools. Get awesome bartenders. Demand great cocktails. Mix it up based on context.
- 3. 90% of theworld’s data was generated over the last 2 years…
- 4. That changes everything: howwe collect, store, manage, analyze and visualize data.
- 5. Data > Information> Knowledge > Wisdom Past | Present | Future
- 6. Big Data isso 2012 Cloud Social the path of the enterprise Mobile Big Data Algorithms next disruption? Dion Hinchcliffe. Dachis Group.
- 7. Algorithms • Machine learning –Predictions – Clustering • Statistical models working at scale – – – – Counting Comparing Ranking Filtering
- 8. Geeky is thenew ‘sexy’ * *
- 9. Geeky? Yup … • • • • • • • • • • Java,R, Python SQL, RDBMS, DW, OLAP NoSQL, Hbase, Cassandra Hadoop, HDFS, MapReduce & Yarn Pig, Hive, Impala, Shark ETL, Webscrapers, Sqoop, Flume Knime, Weka, RapidMiner SPSS, SAS, OBIEE D3.js, Gephi, Tableau, Flare, Shiny Microsoft Excel
- 10. Data scientists enable thecreation of data products.
- 11. A data productis … • Curated and crafted from raw data • Meshed together from disparate sources, some with structured and some with unstructured data • A result of exploration and iterations • Answers known unknowns, or unknown unknowns • Triggers immediate business value • A probabilistic window of future events or behavior
- 12. Financial services isthe world’s most heavily regulated industry.
- 13. Risk is uncertaintyabout a future outcome.
- 14. Key Risk Indicator(KRI) is a management measure used to detect an adverse impact or prevent the possibility of future adverse impact. Expressed as a data product.
- 15. Compliance risk isthe current and prospective risk to earnings or capital arising from violations of, or nonconformance with, laws, rules, regulations, prescribed practices, internal policies, and procedures, or ethical standards. This risk exposes the institution to diminished reputation, fines, civil money penalties, payment of damages, and the voiding of contracts.
- 16. In early September2011, the Swiss bank UBS announced that it had lost over 2 billion dollars, as a result of unauthorized trading performed by Kweku Adoboli, a director of the bank's Global Synthetic Equities Trading team in London. In April and May 2012 large trading losses occurred at JPMorgan's Chief Investment Office, based on transactions booked through its London branch. Trader Bruno Iksil, nicknamed the London Whale, accumulated unauthorized outsized CDS positions in the market. The original estimated trading loss of $2 billion was announced, with the final actual loss expected to be substantially larger. HSBC Holdings Plc agreed to pay a record $1.92 billion in fines to U.S. authorities for allowing itself to be used to launder a river of drug money flowing out of Mexico and other banking lapses. is the world’s most heavily regulated industry In January 2008, the bank Société Générale lost approximately €4.9 billion closing out positions over three days of trading beginning January 21, 2008, The bank states these positions were fraudulent transactions created by Jérôme Kerviel, a trader with the company. After analyzing post-loss & causal factors, they all had a good chance of being prevented or detected if Key Risk Indicators (KRIs) had provided information that could be aggregated, analyzed, and escalated. 16 DIGITAL REASONING | CONFIDENTIAL
- 17. Example: UBS RogueTrading 2012
- 18. Example: UBS RogueTrading 2012
- 19. Many KRIs definedto monitor trading risk
- 20. With access tothe right data product – We can build an “Holistic view” of a Trader’s risk profile
- 21. Human risks arehard to predict: Even the best designed risk controls are subject to the failings of people’s experience, attitude, mindset and values.
- 22. Traders are people. Peoplecommunicate with people. People communicate using human language.
- 23. Human language isa rich data source that enables data scientists to study people’s past behavior or predict future behavior.
- 24. Human language isdirty data. Different languages. Full of ambiguity. Large amounts of it, and very noisy. Difficult to count things.
- 25. "You shall knowa word by the company it keeps." - J. R. Firth, English linguist
- 26. Making human languagetractable: Resolving entities, facts & relationships In time and space *09/26/201 3 *Social Interaction *09/26/201 3 26
- 27. Transforming data intoknowledge 26 Sept VZ (*Verizon? ) Tom Watson Social Interaction UBS +44-20-7567-8000 *Social Hans Interaction Gruber 1 Finsbury Avenue London, UK EC2M 2PP hans.gruber@uk.ubs .com Kurt Dyson kdyson@richardson
- 28. Combining data frommultiple sources – Social media 28 28
- 29. Combining data frommultiple sources – Financial system 29
- 30. Combining data frommultiple sources – Trade Surveillance On September 29, UBS trader Hans Gruber executes a short on 100K shares of AAPL shares for Kurt Dyson, a high profile buy side client of the firm. On Oct. 7, Apple announces disappointing iPhone6 sales resulting in a 10% share price drop and a windfall profit for Kurt Dyson based on the Gruber’s short order.
- 31. Data Cocktail –Holistic View of Hans Gruber 26 Sept VZ Verizon (*Verizo n?) Tom Watson Social Interaction UBS +44-20-7567-8000 Hans Gruber 1 Finsbury Avenue London, UK EC2M 2PP hans.gruber@uk.ubs .com Kurt Dyson kdyson@richardson 31
- 32. Analyzing Enron’s publicemail data. Bill DiPietro & Jascha Swisher Digital Reasoning.
- 33. Example: KRI forHuman Language “Legal Entity” on restricted trading list occurring in electronic communications + Legal Entity occurring in the context of “deal related” language + Communication “outside” company firewall
- 35. PHONE (615) 370-1860 EMAILmarten@digitalreasoning.com WEB digitalreasoning.com * *
Editor's Notes
- #6 "Knowledge is a process of piling up facts; wisdom lies in their simplification." --Martin Fischer
- #15 A Key Risk Indicator, also known as a KRI, is a measure used in management to indicatehow: risky an activity is to detect an adverse impact or prevent the possibility of future adverse impact
- #17 So how serious of a problem has this become for the commercial sector? Let’s take a look at some recent cases involving top tier financial institutions.In September 2011, the Swiss bank UBS announced that it had lost over 2 billion dollars, as a result of unauthorized trading.In this case the trader was actually talking about the unauthorized trading in e-mail, specifically he was using the terms “umbrella” and “slush funds” in communication with colleagues.There was also a chat transcript in which a trade support analyst told Adoboli to cancel and rebook a trade to change the settlement date. The trade support analyst must have known it was a fake trade since If it were a genuine transaction, you couldn’t rebook and just move the settlement date. In April and May 2012 large trading losses occurred at JPMorgan's Chief Investment Office, based on transactions booked through its London branch. This is the famous case, nicknamed the London Whale, where a rogue trader had accumulated unauthorized outsized positions in the market.Again there were e-mails with damaging information, JPMorgan traders reportedly called regulators from the Office of the Comptroller of the Currency "stupid," and other traders fretted in an email that "we are going to crash.”At one point Iksilemailed a senior trader advising against increasing the bet, as the size of the trades were becoming “scary”. Advised take the “full pain” now. HSBC Holdings Plc agreed to pay a record $1.92 billion in fines to U.S. authorities for allowing itself to be used to launder a river of drug money flowing out of Mexico and other banking lapses.HSBC conceded that its anti-money laundering measures were inadequate and that it has taken big steps in beefing up its controls. In return for being spared prosecution, HSBC said it would continue to strengthen its compliance policies and procedures. a Senate investigation concluded that HSBC’s lax controls exposed it to money laundering and terrorist financing. In December 2, 2001, Enron filed for Chapter 11 Bankruptcy, with $63.4 billion in assets it was the largest corporate bankruptcy in U.S. history. Enron’s downfall was a result of its complex financial statements and complex business model. "the primary motivations for Enron's accounting and financial transactions seem to have been to keep reported income and reported cash flow up, asset values inflated, and liabilities off the books.” A lot of these complex financial dealings were discussed to some extent in e-mail(BILL, YOU NEED TO SPEAK TO HOW THIS RELATES TO THIS AUDIENCE – THE GOVT. I SUGGEST YOU SPEAK TO THIS ON THIS SLIDE OR ADD ANOTHER ONE.)
- #18 Copyright Deloitte 2012.
- #19 Copyright Deloitte 2012.
- #20 Copyright Deloitte 2012.
- #21 Copyright Deloitte 2012.
- #26 Machines can help data scientists. Predict meaning of words and phrases based on context.