Systems and user interfaces for holistic, data-driven investigation of bad actor behavior based on clustering and scoring of related data

• US 10,223,748 B2
• Filed: 08/17/2016
• Issued: 03/05/2019
• Est. Priority Date: 07/30/2015
• Status: Active Grant

• Current Assignee: Palantir Technologies Incorporated
• Original Assignee: Palantir Technologies Incorporated
• Inventors: Hunter, Sean, Kumar, Aditya, Albertson, Jacob
• Primary Examiner(s): Akintola, Olabode
• Application Number:
  US15/239,482
• Publication Number:
  US 20170032463A1
• Time in Patent Office: 930 Days
• Field of Search: None
• CPC Class Codes: G06F 16/2457 with adaptation to user needs
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 3/0482 Interaction with lists of selectable items, e.g. menus
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange

Abstract

Embodiments of the present disclosure relate to a data analysis system that may automatically generate memory-efficient clustered data structures, automatically analyze those clustered data structures, automatically tag and group those clustered data structures, and provide results of the automated analysis and grouping in an optimized way to an analyst. The automated analysis of the clustered data structures (also referred to herein as data clusters) may include an automated application of various criteria, rules, indicators, or scenarios so as to generate scores, reports, alerts, or conclusions that the analyst may quickly and efficiently use to evaluate the groups of data clusters. In particular, the groups of data clusters may be dynamically re-grouped and/or filtered in an interactive user interface so as to enable an analyst to quickly navigate among information associated with various groups of data clusters and efficiently evaluate those data clusters in the context of, for example, a risky trading investigation.

Claims

1. A computer system comprising:

one or more computer readable storage devices configured to store;

a plurality of computer executable instructions;

at least one data cluster stored in a memory of the computer system, wherein the at least one data cluster is associated with a data clustering strategy and is generated according to the data clustering strategy, the data cluster including at least;

a plurality of trade data items including information associated with trades of a trader;

a plurality of external event data items including information associated with at least one of a trade confirmation, a trade settlement, an exchange margining, or a cash flow associated with a trade;

a plurality of logical connections among the data items in the data cluster, wherein each logical connection indicates a relationship between at least two of the data items; and

wherein all the data items in the data cluster are linked with one another, either directly or indirectly, by the logical connections;

a trading risk indicator configured to utilize at least a subset of the data items in the data cluster; and

one or more hardware computer processors in communication with the one or more computer readable storage devices and configured to execute the plurality of computer executable instructions in order to cause the computer system to;

access the data cluster, including the plurality of trade data items and the plurality of external event data items, from the one or more computer readable storage devices;

access the trading risk indicator from the one or more computer readable storage devices;

apply the trading risk indicator to the data cluster by;

analyzing the plurality of trade data items; and

analyzing the plurality of external event data items to identify a subset of external event data items; and

generate an alert in response to determining that the subset of external event data items includes more external event data items than a threshold number of external event data items.

2. The computer system of claim 1, wherein the at least one data cluster is further associated with data cluster analysis rules and/or data cluster scoring rules that comprise the trading risk indicator, wherein the trading risk indicator is observable or computable from the data items in the particular data cluster.

3. The computer system of claim 2, wherein the trading risk indicator is configured for indicating at least one of heightened risk or reduced risk.

4. The computer system of claim 1, wherein the plurality of computer executable instructions further cause the computer system to generate user interface data for rendering an interactive user interface on a computing device, the interactive user interface including one or more selectable elements useable by a user for indicating the generated alert.

5. The computer system of claim 4, wherein the plurality of computer executable instructions further cause the computer system to use at least the generated alert to generate a summary report based on the data cluster analysis rules and/or the data cluster scoring rules.

6. The computer system of claim 5, wherein the plurality of computer executable instructions further cause the computer system to receive feedback from the user through the interactive user interface, the feedback containing a suggestion for improving the summary report generated based on the data cluster analysis rules and/or the data cluster scoring rules.

7. The computer system of claim 6, wherein the plurality of computer executable instructions further cause the computer system to update the data cluster analysis rules and/or data cluster scoring rules based on the feedback received from the user for improving the generated summary report.

8. The computer system of claim 1, wherein the trading risk indicator is a possible dummy trade indicator for identifying when a trade is cancelled or amended before an external event that might affirm the trade is real, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify cancelled or amended trades of the trader;

determining a cancellation or amendment time associated with each cancelled or amended trade of the trader;

analyzing the plurality of external event data items to identify an external event associated with each cancelled or amended trade; and

determining the time associated with the external event associated with each cancelled or amended trade;

wherein the alert is also generated in response to determining that the cancellation or amendment time for each cancelled or amended trade is prior to the external event time associated with the cancelled or amended trade.

9. The computer system of claim 1, wherein the trading risk indicator is part of a trading when absent scenario for detecting whether a person'"'"'s trading activity coincides with unusual patterns in security badge data, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify the trades performed by the trader;

determining an execution time associated with each trade performed by the trader;

analyzing the plurality of external event data items to identify time windows the trader is not in a building based on security badge usage data for the building; and

determining a subset of external event data items, wherein each external event data item of the subset of external event data items is associated with a respective execution time of one of the trades performed by the trader and a time window for which the trader is not in the building.

10. The computer system of claim 9, wherein the security badge usage data comprises at least one of:

time windows the trader is in the building;

time windows the trader is not in the building;

times the trader used a security badge to enter the building; and

times the trader used a security badge to leave the building.

11. The computer system of claim 1, wherein the trading risk indicator is part of a suspicious badge activity scenario for detecting whether a trader'"'"'s trading activity coincides with unusual patterns in security badge data, and applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify the trades performed by the trader;

determining an execution time associated with each trade performed by the trader;

analyzing the plurality of external event data items to identify time windows the trader is not in a building from the security badge usage data for the building;

determining unusual time windows from the time windows the trader is in the building based on unusual security badge usage patterns; and

determining a subset of external event data items, wherein each external event data item of the subset of external data items is associated with a respective execution time of one of the trades performed by the trader and an unusual time window for which the trader is in the building.

12. A computer system comprising:

one or more computer readable storage devices configured to store;

a plurality of computer executable instructions;

at least one data cluster stored in a memory of the computer system, wherein the at least one data cluster is associated with a data clustering strategy and is generated according to the data clustering strategy, the data cluster including at least;

a plurality of trade data items including information associated with trades of a trader;

a plurality of profit and loss (PNL) data items including information associated with the trader'"'"'s PNL over a time period; and

a plurality of logical connections among the data items in the data cluster, wherein each logical connection indicates a relationship between at least two of the data items; and

wherein all the data items in the data cluster are linked with one another, either directly or indirectly, by the logical connections;

a trading risk indicator configured to utilize at least a subset of the data items in the data cluster; and

one or more hardware computer processors in communication with the one or more computer readable storage devices and configured to execute the plurality of computer executable instructions in order to cause the computer system to;

access the data cluster, including the plurality of trade data items and the plurality of PNL data items, from the one or more computer readable storage devices;

access the trading risk indicator from the one or more computer readable storage devices;

apply the trading risk indicator to the data cluster by;

analyzing the plurality of trade data items; and

analyzing the plurality of PNL data items; and

generate an alert in response to application of the trading risk indicator to the data cluster.

13. The computer system of claim 12, wherein the trading risk indicator is a PNL smoothing indicator for detecting whether a trader'"'"'s PNL has an unreasonably smooth volatility profile given price volatility of asset classes traded and reflected in the PNL, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of PNL data items to identify the asset classes traded by the trader and reflected in the trader'"'"'s PNL over the time period;

analyzing the plurality of trade data items to determine volatility of each of the asset classes traded by the trader and reflected in the trader'"'"'s PNL over the time period;

determining volatility of the trader'"'"'s PNL over the time period;

determining a volatility discrepancy between the trader'"'"'s PNL and the asset classes by comparing the volatility of the trader'"'"'s PNL over the time period to the volatility of each of the asset classes traded by the trader and reflected in the trader'"'"'s PNL over the time period; and

generating the alert in response to determining the volatility discrepancy between the trader'"'"'s PNL and the asset classes is greater than a threshold of volatility discrepancy between the trader'"'"'s PNL and the asset classes.

14. The computer system of claim 12, wherein the one or more computer readable storage devices are further configured to store a plurality of trading risk scenarios for detecting, confirming, or denying a particular hypothesis, and wherein each trading risk scenario comprises one or more trading indicators.

15. The computer system of claim 12, wherein the trading risk indicator is configured for indicating at least one of heightened risk or reduced risk, and wherein at least one of the plurality of trading risk indicators is observable or computable from the data items in the particular data cluster.

16. The computer system of claim 12, wherein the data cluster further includes a plurality of account data items including information associated with trading accounts associated with the trades of the trader, wherein the trading risk indicator comprises an internal transfer with non-zero PNL indicator for identifying trades performed by a trader that are internal trades between internal trading books or internal trading accounts and also associated with a non-zero PNL, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify trades of the trader;

analyzing the plurality of account data items to identify trading accounts associated with the trades of the trader;

determining which trades of the trader has an internal counterparty using the identified trading accounts associated with each trade of the trader;

for each trade determined to have an internal counterparty, determining whether that trade has a non-zero PNL using the plurality of trade data items; and

generating the alert in response to determining that a trade has both an internal counterparty and a non-zero PNL.

17. The computer system of claim 12, wherein the data cluster further includes a plurality of account data items including information associated with trading accounts associated with the trades of the trader, wherein the trading risk indicator comprises a mark arbitrage indicator for identifying trades performed by a trader that are internal trades between internal counterparties that have different accounting regions, wherein the internal counterparties comprise internal trading books or internal trading accounts;

and wherein applying the trading risk indicator further comprises;

analyzing the plurality of trade data items to identify trades of the trader;

analyzing the plurality of account data items to identify trading accounts associated with the trades of the trader;

determining which trades of the trader has internal counterparties using the identified trading accounts associated with each trade of the trader;

for each trade determined to have internal counterparties, determining whether the internal counterparties have different accounting regimes using the plurality of account data items; and

generating the alert in response to determining that a trade has internal counterparties with different account regimes.

18. The computer system of claim 12, wherein the data cluster further includes a plurality of external event data items including information associated with a compensation cycle for the trader, wherein the trading risk indicator comprises a PNL jump in line with compensation cycle indicator for identifying big, positive PNL changes highly correlated to timing of a compensation cycle, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of PNL data items to identify a big, positive change to the trader'"'"'s PNL;

determining a time associated with the big, positive change to the trader'"'"'s PNL;

analyzing the plurality of external event data items to identify a compensation cycle;

determining a time associated with a change in the compensation cycle; and

generating the alert in response to determining that the time associated with the big, positive change to the trader'"'"'s PNL is within a threshold time period right before the time associated with a change in the compensation cycle.

19. The computer system of claim 12, wherein the trading risk indicator comprises an adverse execution selection indicator for identifying trades booked away from the market in order to create PNL or cash, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify trades of the trader;

determining an actual execution cost associated with each trade of the trader;

determining an expected execution cost associated with each trade of the trader; and

generating the alert in response to determining the deviation between the expected transaction cost and the actual execution cost of each trade of the trader is above a threshold.

20. The computer system of claim 12, wherein the trading risk indicator comprises a trading away from bucket VWAP indicator for identifying trades that are executed at a price which should on average be attainable by a trader, and wherein applying the trading risk indicator further comprises:

analyzing the plurality of trade data items to identify all trades by the trader over a time period;

determining a traded product associated with a trade by the trader over the time period;

determining an execution price associated with a trade by the trader over the time period;

analyzing the plurality of trade data items to identify all trades in a firm over the time period associated with the traded product or a similarly traded product, wherein the similarly traded product has a higher than average correlation to the traded product;

determining an execution price associated with each trade in the firm over the time period associated with the traded product or a similarly traded product;

determining the volume weighted average price (VWAP) of the execution prices associated with all trades in a firm over the time period associated with the traded product or the similarly traded product; and

generating the alert in response to determining the execution price associated with a trade by the trader over the time period deviates from the VWAP by a threshold.