Flared graphs on a Mexican flag

Predict the unpredictable

Data intelligence is much more than just aggregated statistics: It is the analysis of external and operating data on a regular basis that puts decision makers and customers in the driving seat. And it opens up new possibilities for making predictions and simulations about future events more reliable, helping central banks, for example, with disaster management and cash supply before, during, and after a crisis.

There is no doubt about it: In terms of safety it would be a huge plus, even a life-saving measure, if natural events such as earthquakes, hurricanes, or tsunamis could be reliably predicted. Nature, however, is difficult to calculate. Geoscientists can recognize factors that indicate an increased probability of e.g. earthquakes, but this is still nothing more than a probability. Precise predictions for the location or timing of tectonic movements and their consequences are still wishful thinking.

Illustration of a world map with many bright dots and notes

Maybe data and machines could help? There are numerous approaches, all trying to make the unpredictable predictable. If a crisis has already occurred, and the rapid flow of events, of information and data, only has to be skimmed off and evaluated, the process could be relatively simple. AIDR, for example, short for “Artificial Intelligence for Disaster Response”, works according to this principle: The free and open source software automatically collects and classifies tweets that are posted during humanitarian crises. “Far too much data is produced via social media during crisis situations for humans to manage on their own,” explains Muhammad Imran, Research Scientist at the Qatar Computing Research Institute (QCRI) and one of the eight collaborators in the project.

Current and historical facts make predictions possible

In addition, the data is too rich and complex for machines to successfully process. “AIDR uses the best of both worlds by combining human and machine intelligence,” Imran continues. At first, the software collects tweets related to a disaster, filtering them using keywords or hashtags like “hurricane” and “#Sandy”. This stage is just like a regular search on twitter.com and the resulting tweets will contain the keywords or hashtags, but may not be relevant to disaster response or to the specific information needs of humanitarian organizations. However, a filter called “Tagger” then classifies tweets by topics of interest, such as “Infrastructure Damage” and “Donations”. This classification is performed automatically and then visualized based on a set of human-tagged items, provided through MicroMappers.org, a joint initiative with the United Nations.

To be clear: The AIDR concept consists of managing information during a crisis, more or less like a look in the rearview mirror. It analyzes the past and the present, though admittedly these might still allow it to foresee developments over the hours or days to come. Beyond this, predictive analytics encompasses a variety of statistical techniques – ranging from predictive modelling to machine learning and data mining – that analyzes current and historical facts to make predictions about future events possible and more reliable.

Algorithms put decision makers in the driving seat

In a laboratory setting, scientists Bertrand Rouet-Leduc, Claudia Hulbert, Nicholas Lubbers, and others have shown that it is possible to predict “labquakes” by applying new developments in machine learning (ML), a technique that exploits computer programs that expand and revise themselves based on new data. “We use ML to identify telltale sounds – much like a squeaky door – that predict when a quake will occur,” they write. By listening to the acoustic signal emitted by a laboratory fault, machine learning can predict the time remaining before the fault fails with great accuracy, the authors continue.

As the algorithm is capable of spotting crucial trends in the data, spots that are often missed by recordings in the real world, this approach could be applied to predict avalanches, landslides, and machine part failures. This puts decision makers in the driving seat – not simply staring at the rearview mirror, but drawing conclusions from correlations, creating transparency, and forecasting the future.

Logo of the Banco de Mexico

An example of how data intelligence can support disaster management in real life is shown by an ingenious plan designed in 2017 by Banco de México to support the continuity of cash distribution in any specific part of the country affected by natural catastrophes. The plan is aimed at securing the supply of banknotes in places where the banking infrastructure has collapsed – as cash is the only form of payment that can be used immediately in times of crisis.

Data intelligence is changing crisis management dramatically

Maintaining banknote supply when infrastructures collapse

Operated by Banco de México, Mexico’s Banking Association, Banjercito (a bank specifically created to provide banking services for the Army and Navy), the Ministry of Finance, and supported by the Army and Navy, the plan consists of four stages and runs until services are normalized. “When three out of four bank branches in any particular Mexican city are unable to operate, the assistance phase of the plan is activated,” says Alejandro Alegre Rabiela, General Director of Currency Issuance and Main Cashier at Banco de México. “Banco de México, the Banking Association, and Banjercito will do their best to install cash service modules in no more than 24 hours,” Alegre continues.

Cash Service Modules and personnel

Banjercito and Mexico’s Banking Association determine the number of cash service modules to be brought to the affected region, the quantities and denominations to be sent, and the maximum amount allowed per transaction at each module. Banjercito notifies Mexico’s Banking Association, the Ministry of Finance, and Banco de Mexico when the Cash Service Modules start operation and informs them of their locations, so that they can arrange for the local media of the affected region to publish the information. Residents of the region are subsequently informed about the module locations, the maximum amount that can be withdrawn per transaction, the type of valid cards, and any other valuable information regarding the operation of the cash service modules.

Shipment of Cash Service Modules

Banco de México will transport, in one or more aircrafts, the banknotes, the cash service modules, and Banjercito’s personnel that will operate the equipment,” Alegre adds. Should it not be possible to transport the equipment and personnel in Banco de México’s aircrafts, the Army or Navy will provide logistical support to install the modules, together with custody services for personnel and cash. While the plan is being implemented, banks will not charge customers fees for withdrawing banknotes. Additionally, Banjercito will inform all participating partners on a daily basis about the operation of the modules, as well as the amount withdrawn per denomination.

In return, Mexico’s Banking Association will provide daily information regarding the progressive restoration of banking services at bank branches and ATMs. When the banking infrastructure is functioning at least at 50 percent of its standard capacity, the assistance phase is over, and the cash modules are removed from the sites. The plan has already been successfully activated after two earthquakes occurred in September 2017, causing major damage in different Mexican States: the first affecting Oaxaca and Chiapas; and the second, Mexico City, the State of Mexico, Puebla, and Morelos.

Real-time insights through a combination of analogue and digital

Specifically, the plan was put into operation during October and November 2017, in Oaxaca, where damage to banking infrastructure was severe. Thanks to this plan, cash requirements for 114.7 million pesos or 5.8 million US dollars could be satisfied.

“In general, data intelligence opens up new possibilities for disaster management before, during, and after a crisis,” says Dr. Marcus Schmeisser, Product Manager Compass Banknote Intelligence. “It is changing humanitarian operations and crisis management dramatically.” For example, data intelligence can help to define the required locations and numbers of cash service modules in advance, based on predicted cash demand. Additionally, it can assist with the transport and logistics of equipment, personnel, and banknotes to ensure the highest possible reliability and efficiency.

Data intelligence opens up new possibilities for disaster management – before, during, and after a crisis.

Dr. Marcus Schmeisser, Product Manager Compass Banknote Intelligence

Regarding the cash supply after natural events like the earthquake in Mexico, data intelligence can be one way to handle all information and derive the right decisions: “Here, data intelligence not only underlines the importance of always having a clear view of the number of banknotes that are in circulation, but also allows operators to point all decisions in the right direction based on key data and facts – be it in Mexico or elsewhere,” Schmeisser adds.

Related topics

Virtual dashboard with analytical graphs

Compass Banknote Intelligence

More transparency, more insights: The combination of analogue processes with digital solutions offers real-time insights into cash processing and the properties of the banknotes. Learn more how Compass Banknote Intelligence delivers all the information central banks need to make their cash cycles secure, efficient and digitally mapped.

READ MORE

Get in touch

If you have any questions about our end-to-end business solutions or about our company, seek expert advice, or want to give us your feedback, our team is here to support you, anytime.