Alloscope, a tool to help those in vulnerable situations

Could it be that analysing a person's call records may reveal unusual behaviour? This theory, developed by a team of researchers at Orange, is set to enhance the range of medical and social monitoring tools. Known as Alloscope, the project will be demonstrated at the 2019 Research Exhibition.

This research work was recompensed by the Grenoble Faculty of Medicine at the 2018 Medical Research Day.

“Tell me who you call and I’ll tell you how you are”. This phrase, borrowing heavily from a well-known proverb, is the watchword for Alloscope, a research project led by a team of scientists from Orange. Targeting professionals in the medical and social sector, this project aims to analyse the telecommunications patterns of people in vulnerable situations that reflect changes in their behaviour. An illustrative demonstration will be presented at the 2019 Research Exhibition.

Your phone, the first port of call

In France, a grid called AGGIR is used to measure the degree of a person’s loss of autonomy. Among the activities assessed, the loss of ability to use a telephone (and thus to alert as required) is a sign of significant loss of autonomy. This encouraged Hervé Provost’s team to get their thinking caps on. As Hervé, a project manager at Orange, explains: “We realised that telephone usage was in fact an excellent way to measure variations in a person’s condition (for instance, social links, psychological status, or physical shape).What’s more, it has the great advantage of being part of life on a large scale in almost every sector and geographic region. While environmental or wearable sensors are being studied to monitor the elderly, among others, our researchers realised that Orange’s information system was already a huge source of information.”

Telecom indicators

When looking at call records (CRA), in other words itemised billing, they found that their analysis elicited very valuable behavioural data. Traces from the 3 network types (RTC, ToIP and mobile) contain time-based items (call time and duration), social items (call direction and number of different numbers dialled or received), and spatial items (the various antennas used during mobile calls). Thus, incoming and outgoing telephone calls show changes in the intensity and diversity of the individual’s social links. “Of course this is only data from operators! It does not include any content. All those wishing to benefit from this scheme must give their consent”, adds Hervé Provost. Using this information, professionals would be in a better position to monitor vulnerable persons in the long run, as well as the effects of certain treatments, for example.

Monitoring – from individual to populace?

Another observation is the correlation between circadian rhythm and telephone usage. In a study recompensed by the Grenoble Faculty of Medicine at the 2018 Medical Research Day, Timothée Aubourg, a PhD student working at Orange, noted that it was worthwhile studying telephone activity over a 24-hour period. Any asymmetry between incoming and outgoing calls on a one-day scale provides an additional indicator of a person’s psychological status (more specifically, depression). Once all the indicators have been analysed, “an AI algorithm can be set up” explains the mathematician. This may well enhance the “digital phenotype” of the person monitored.

Though the project is still at the research stage, its large-scale deployment is already under consideration. In fact, by broadening the process to include the inhabitants of a city, for instance, the indicators collected could in turn act as a barometer. Another way to gauge the famed mood of French people. Using Big Data technology, there is no need for identification or consent, as the analysis would be completely anonymous.

Read also on Hello Future

FoodProX uses machine learning to detect ultra-processed food


Data from mobile phones shed light on COVID-19 lockdowns in France


Orange 2021 Research Exhibition: let’s prepare for the future!


Measuring network usage for each mobile application


A machine that lends an ear


The Internet of Things (IoT) meeting the needs of trading ports


Federated AI and reactivity of network alert systems


2030 outlook: which networks for a world in transition?