Can large data prevent Ebola?

The Ebola virus is the devil, will let the patient spit visceral tissue, qiqiao bleeding and death, the death rate is as high as 70%. There are reports that Ebola will log in to China at the end of the month, and what can the big data analysis do in the spread of Ebola? Here is a BBC article for domestic colleagues to learn from.

For public health experts, it is their greatest challenge to popularize the way and means of spreading the virus to the public.

The Ebola virus that broke out in West Africa has so far claimed more than 4,000 lives. Can large data analysis help when emergency rescue teams, medical charities and NGOs struggle to control the virus? More and more data scientists give positive answers.

Move Map

In a region where other reliable resources are scarce, mobile phones are proving to be a reliable source of large amounts of data, since mobile phones are widely available even in Africa's poorest countries. After handing over anonymous voice and text messages from 150,000 handsets in Senegal to Flomind (Flowminder, a Swedish non-profit organization), the company Flomind a detailed map of the region's typical population migrations. "

The authorities can then see where the best places to set up treatment centers are, and where to limit travel is the most effective way to control the epidemic. The drawback of these data, however, is that they are historical data, and what the authorities need is to be able to reflect the flow of the population in real time, because the migration of the population will change according to the change of the epidemic.

That's why the CDC is also collecting mobile data from mobile operators to find out where the number of helpline is greatest.

Mobile phone data in West Africa are being used to map population migration and to predict the possible ways and means of transmission of the Ebola virus.

A surge in the number of calls to a regional helpline would mean an outbreak, which would also alert the authorities to mobilizing more resources to the area.

The Map software company Eisre (ESRI) is helping the Centers for Disease Control and Prevention (CDC) visualize this data and overwrite existing census data sources to create a richer visual image. Depending on the level of activity of the mobile base station, the hotspot maps they draw can clearly reflect where people are, and more crucially, where they are migrating and how far they have migrated.

"We've never had such a large, anonymous cell phone data as a kind of information before," said Nuria Oliver, Nureoliver's science director at Telefónica. "The most positive effect is that we can help emergency agencies and the government predict how the disease may spread."

"Before that, they had to rely on sporadic information, such as field investigations, police and hospital reports." "Mobile phones have again proved to be an ideal and effective way to deliver health messages," he said.

Lessons from Cholera

This analysis of mobile data has been successfully applied to other health crisis events.

For example, in 2010, after the earthquake in Haiti, a joint study team from the Swedish Caroline College and Columbia University analyzed data from 2 million mobile phone calls from the Caribbean operator Digicel Haiti Network.

Following the Haiti earthquake disaster of 2009, Digicel's mobile data were used to track population movements.

This has enabled the United Nations and other humanitarian agencies to understand and control population movements in disaster relief operations and subsequent cholera outbreaks, which means that they can allocate resources more effectively and identify areas of increased risk in the new cholera outbreak.

Data analysis of 15 million mobile phones is also used to reflect and predict the spread of malaria in Kenya.

But Ms Oliver admits: "Moving data can only help us understand a part of the situation." “

Can large data analysis provide a measure of effectiveness?

Experts believe that in order to gain a more comprehensive understanding, we need more data sources and the ability to quickly analyze the data.

"Large data analysis is about pooling many different data sources and mining data to find patterns." "Our data comes from health clinics and doctors ' reports, media reports, social media commentary, information received by public health workers on the ground, data from retailers and pharmacies, travel ticket purchases," said Frances Dare, Francis Dayer's health consulting director. Hotline data, as well as geo-spatial tracking data. “

British hospital vessel Rfaargus ready to ship supplies to Sierra Leone

Peter Jung, chief technology officer at BAE Systems, the Peder Jungck, said such an analysis could also be used to measure the containment epidemic, and whether education campaigns and treatments were effective.

"For example, doctors can tell how many people are taking appropriate precautions to reduce the spread of disease by analyzing social-media data sets for high-risk populations, and the proportion of people who ignore preventive reminders. For Ebola, analysts who study large datasets can also analyze potential health challenges, as well as local environmental factors such as whether the weather has an impact on the speed of disease transmission.

Cross-border communication

In this era of global travel, it is easier to spread the virus across borders than ever before, especially for Ebola, a virus with a 21-day incubation period. Both Europe and the United States are on high alert and are being cleared at some airports.

But in the digital age at least, it is much easier to track down potential carriers of the virus than ever before. "Ports, train stations and airport data, as well as license plate recognition techniques, can help track potential virus carriers and determine who they may have contacted." Davidbolton, head of the medical department at Qlik, a big data analyst, said. He has developed an Ebola tracking app.

Since Tuesday, Heathrow Airport has begun screening passengers with Ebola-related symptoms.

Predicting epidemic trends through social media

Through social media and dynamic data on search engines, analysts are predicting that epidemic trends are also getting better.

While the Google flu trend, which predicts the likelihood of a possible cold virus outbreak through the frequency of key search terms used by people, has proved unreliable, other ways of using a wider dataset have been more successful.

For example, Accenture, a business consultancy, the large data specialist SAS and the University of North Carolina, said they predicted that the 2012-13 flu season in the United States would be released three months ahead of the CDC's official warning.

"By analyzing social media, such as blogs, web forums and Twitter, we can find early warning signs of public security events." Frances Dare of Accenture said.

"In 2013, we reduced the number of keywords that showed flu symptoms to 152, labeled where the keywords were being used and predicted a flu outbreak two months before official data were released," he said. ”

Looking for a cure?

Tim Gamble, chief consultant at Datamonitor Healthcare, believes that large data analysis can also help understand viral genetics, such as why certain types of viruses are more lethal and why some people are more resistant to viruses. "The treatment of HIV retrovirus has not occurred before many people begin to die of AIDS," Pfizer, an American pharmaceutical company that specializes in infectious disease research. I was involved in Pfizer's HIV product development work, and we found that some people in Scandinavia were more resistant to HIV than others. ”

The Ebola virus is highly contagious and spreads through blood, vomit and feces.

"So we developed a drug that mimics the mechanisms that people are fighting against the HIV virus," he said. He believes the same method applies to the Ebola virus.

In short, large data analysis has been applied to various levels in the fight against Ebola. But as Qlick's David Burton says, "We're still learning from the beginning, we've never had this level of data." It is premature to assess whether large data analysis has a significant impact on disease diagnosis and transmission, but at least large data analysis can help us decide how to allocate our resources. ”

Original address: http://www.bbc.com/news/business-29617831

BBC News-Journalist: Matthew? Wall

Translation: Wang Changan, Jia Wenjing, Furo

Proofreading: Chang

Introduction to Ebola virus

Ebola (Ebola virus) also translated the Ebola virus. is a very rare virus, in the 1976 in southern Sudan and the Congo (former Zaire) (old) in the Ebola River region found its existence, caused widespread concern and attention of the medical community, "Ebola" from this name. is a generic term used to address a group of viruses belonging to the fiber virus branch Ebola virus. It is a potent infectious virus that causes Ebola haemorrhagic fever in humans and primates, with a high mortality rate between 50% and 90%, mainly for stroke, myocardial infarction, low blood volume shock or multiple organ failure.

The Ebola virus is a potent virus that causes Ebola haemorrhagic fever in humans and primates, the resulting Ebola haemorrhagic fever (EBHF) is the world's deadliest viral hemorrhagic fever, with symptoms similar to those associated with the Marburg virus, including nausea, vomiting, diarrhea, skin changes, body aches, Bleeding in vivo, bleeding in vitro, fever, etc.

Ebola virus, biosafety grade 4 (AIDS is 3, SARS is 3, the greater the level of protection more stringent). Virus latency can reach 2-21 days, but usually only 5 days to 10 days.

(Responsible editor: Lvguang)