Oct 27, 2015

The five witnesses responsible for the creation of call sequence tables

On 20-22 July the Prosecution presented the evidence of five members of the Prosecution who were responsible for the creation of call sequence tables (CSTs). The telephone numbers in the CSTs include the numbers of the various networks that, according to the Prosecution, were used by the accused in preparation of the bomb attack. These five witnesses all produced very similar evidence.

The Prosecution explains in court that a CST “presents a chronological and complete sequence of calls relating to a particular phone number referred to as the target number over a specified period of time. The CSTs detail each call including the other telephone number the target number was in contact with; the date and time of the call; the type of call, either voice or short message service, or SMS; its duration; information on the handset used by the target number; and cell information on the cell sector used by the target number at the start of the call. The CST occasionally also may include information on the cell sector used at the end of the call, which is referred to as end cell data.”

There are also CSTs for SMS (text messages), which present the SMS content sent and received by a phone over a relevant data range in a consistent accessible format. The SMS CST provides the data and time of the SMS; the other telephone number involved; the direction of the SMS and its content.

The Chamber, in its decision of 6 May in relation to call sequence tables, witness statements and the legality of the transfer of call data records to UNIIIC and the Prosecution, has deferred its decision on the admissibility of CSTs (and related witness statements) until at least one witness has testified about the provenance of the call data records and the production of the CSTs.

Ms. Kei Kamei

The first witness, Ms. Kei Kamei, has been working as an analyst with the Office of the Prosecutor since 9 November 2009, and previously worked for UNIIIC. She has personally produced numerous (SMS) CSTs and coordinated the standardisation of CSTs within the office. Ms. Kamei gives evidence on the source materials she used to produce the CSTs, which are (i) files for one particular phone number or interest; and (ii) bulk data for multiple numbers on a specific day. Ms. Kamei explains the various component parts of call data records. The Prosecution received this data by making a request for assistance to the Lebanese telecommunications service provider in relation to a specific number of interest. The Prosecution requested the bulk call data records for the period September 2004-December 2005 - which was later even extended - that is all phone calls made in Lebanon between these dates. Ms. Kamei further explains the functioning of the SQL database which was used to analyse the bulk data. Further, for the SMS CSTs only bulk data was used, although not all content of the SMS messages appeared to be available in the call data records. In addition, a list with cell tower names received from the telecommunication companies was used to replace the cell IDs in the CSTs.

Ms. Kamei continues to explain how she created the CSTs out of the available data, which involved quite some manual handling. The Judges express their concern as to the accuracy and reliability of the CSTs produced. Ms. Kamei explains the content of various CSTs shown to her. The amount of errors located and the validation and peer review processes are discussed as well.

During cross-examination Mr. Edwards questions the witness about the creation of her statement, and about spot-checks and quality control of the CSTs. If another analyst checks the CST, usually that person would go to the raw source and extract a few, do spot-checks on certain rows.  There was not a formal protocol. No record was kept of the errors that were found. Ms. Kamei admits that it is possible that some of the CSTs contain mistakes that were not detected.

Mr. Andrew Donaldson

The next witness, Mr. Donaldson has been a Prosecution analyst since March 2009. He produced 45 CSTs, and in his statements he describes the components of the CSTs, as well as the sources, and the creation and double-checking of the CSTs. Mr. Donaldson would give his work a 9.9 on a scale of 10 regarding accuracy. The witness estimates that almost all of the CSTs that the Prosecution intends to rely upon as exhibits have been peer reviewed.

Ms. Helena Habraken

Ms. Habraken is the next analyst of the Office of the Prosecutor giving testimony about the CSTs. In her statement of 18 December 2014 she describes having produced 27 CSTs, and she also describes the work methods and sources of the CSTs. Ms. Habraken states that she thinks that the CSTs she made are highly accurate representations of the call data records. According to her, and generally speaking, most errors in the CSTs are formatting errors, when a number would not be properly displayed or a problem in a date or time format.

Mr. Lachlan Christie

Mr. Christie also is an analyst with the Office of the Prosecutor who produced CSTs. He testifies about the same topics as the previous witnesses. Again his confidence level is high that the CSTs he produced are accurate representations of the call data records.

Mr. Christian Carnus

Another Prosecution analyst who produced SMS CSTs, Mr. Carnus describes the method he used to create the SMS CSTs: the SMS contents records were provided by MTC Touch to the Prosecution in the form of large text files for multiple numbers. Mr. Carnus queried the data through the stored procedures in the SQL database and copied the results into an Excel spreadsheet. Mr. Carnus then formatted the SMS CST. Also Mr. Carnus was asked about his confidence level and he is very confident that the SMS CSTs accurately reflect the SQL database. He confirms that the errors he detected were mainly cosmetic in nature. He also confirms that there was no formal peer review process. There was no cross-examination by the Defence.

Ms. Nadine Stanford

Ms. Stanford is another Prosecution analyst and her statement tendered into evidence contains similar evidence as the previous witnesses.

No comments:

Post a Comment