Shoshana Feld-Sobol
02-07-2021
Share this article
COVID-19
Essential Steps in Saving a City – From High Risk to Low Risk in 3 Weeks

Contrary to popular belief, or wishful thinking, and despite it being over a year since coronavirus disrupted our world, the pandemic is only getting worse in most places, not better. As we continue to live in a time when governments, economies, and livelihoods are subject to hospital bed capacity, the status quo of controlling the pandemic isn’t adequate.

In order to control, monitor, and even predict what cities, neighborhoods, and towns will experience another surge of the virus, we fundamentally believe that AI behavioral intelligence is urgently needed.

This is a white paper guide of how – by using data-driven predictions – a city went from a red zone to green city with 72% more efficiency in first responder resource allocation, and how cases per day dropped 360% from when at its peak, in just three weeks. We offer tangible recommendations and in depth information on how our platform was able to achieve the results it did. 

Background

The only way to save the 20+ million people still infected with coronavirus is through the utilization of behavioral intelligence. In this white paper we will demonstrate how behavior intelligence can map where transmission occurs, as well as predict future outbreaks. These insights equip decision makers with data-driven intelligence used to inform policy and cut the chain of infection.

Transmission hotspots in Florida ranked from low transmission (blue) to high (red) 

Since the inception of the coronavirus pandemic, governments have been releasing guidelines and regulations that respond to already sick patients. This approach to controlling the spread is largely ineffective because it doesn’t stop spreading from occurring, instead just provides information on what cities have higher populations of coronavirus positive people. This doesn’t take into account the 10-14 day seeding and spreading that occurs before patients begin to feel symptoms. 

Current guidelines rely on social-distancing, mask wearing, and one’s own discretion when partaking in social gatherings, work gatherings, or when just leaving their home. While this can mitigate the spread of the virus in some capacities, it falls drastically short when it comes to flattening the curve. Cases have continued to climb at a dramatic rate for the past year, with governments forcibly putting their entire economy at a stand still or operate at the expense of safety guidelines. Flattening the curve requires more than reactionary measures to be truly effective. 

Predictions and behavioral projections provide the only viable solution for breaking the chain of infection instead of pursuing a slow, reactionary policy. Despite the deployment of two effective vaccines on the horizon, a myriad of challenges will persist, including the prospect of a rollout that will take months, if not years, to disseminate worldwide. Neura powers Israel’s National Prediction Model, the “Ichilov Model” with an accuracy rate of 92.1% when predicting infection outbreaks 1, 2, and 3 weeks in advance. 

The spread of COVID-19 has wreaked havoc on economies the world over and has done irreparable damage to society. Cities that have become red zones for new cases – an area that has experienced more than 100 new cases per 100,000 people the week before or an area that has reported COVID-19 test positivity rates of higher than 10% – are often financial and social hubs. Their impact on infection rates and economies is immense, yet authorities continue to struggle to identify and address the spread in these epicenters of importance vis-a-vis the COVID-19 fight. There is a geographical gap between transmission occurrence and the physical residence of people with confirmed cases. Therefore, looking at transmission clusters and transmission is key to understanding riskier and safer zones. 

New development

Neura’s model predicts behaviors to help monitor and control the spread of COVID-19. Using human behavior data, to find indicators that predict transmission, Neura’s presents a visualized dashboard that centers on interrupting interpersonal transmission, the protection of high-risk personnel, and communicating critical information to the relevant authorities where there is a hotspot of activity that can trigger an outbreak. The technology includes multiple facets with features including proximity indexes and heat mapping, data insights for high-risk infection areas, displayed with Neura’s dashboard products. This can safely help governments, businesses, and municipalities deal with changing COVID-19 regulations and adjust measures accordingly.


Transmission risk by neighborhood level 

Neura’s behavioral intelligence platform fuses human behavior data with epidemiological data to generate indicators that identify transmission as it occurs in real time. Insights used to determine where and when to deploy additional resources. The signals are based on previous modeling and heat maps, social distancing adherence, density, and overall population behavior. The key insights include visibility into anonymized and aggregated populations that are becoming mobile post-lockdown and location mapping of newly mobile demographics while also maximizing social distancing perimeters and supporting business needs. Neura aims to help protect against future coronavirus europtions by utilizing relevant data to help predict high-risk transmission areas and deploying relevant resources to those areas such as additional testing and regulation reinforcements. 

Challenge

The above image shows a risk score for each zip code based on factors that share the general likelihood of contracting Coronavirus in those neighborhoods. Predictions and risk score allocation are centered around people coming from CVID-19 hotspots, and are based on parameters such as social distancing observance, crowd gathering, density, and where the bulk of the people in those areas are traveling to and from.

With businesses continuing to face mounting losses and school openings further complicating efforts to limit COVID-19 exposure, our municipality client needed an effective solution that struck a balance between livelihoods and lives. Neura had to act on where to apply critical resources, how and what to communicate, and how to get the city safely back to working order, even if it meant adhering to a new normal.

The city had previously taken appropriate action laid out by government officials and implemented measures provided by the government authorities to mitigate and monitor the spread of COVID-19. The typical tools used in most cities, such as screenings, communication, and public safety enforcement were utilized to the best of their ability. Despite these precautions, over 9.2% of the population became infected; a rate that was considered too high. Despite following government guidelines, the city remained a red zone. The regulations weren’t having a large enough impact on infection rates, and the city was determined to find a solution to curb the spread. Below is a step by step guide for how Neura helped get this city out of high risk. 

How it works

Instead of using the three basic tools to monitor and control the spread of the virus, Neura introduced 15 different tools for monitoring hotspots, potential danger zones, super-spreader events, and risky behavioural patterns. Using behavioral intelligence technology gives visibility into key population movement metrics such as if people are staying home or when they are mobile, their distances from homes, and how long they spend out of the home. 

Hours Out of Home

The image above shows the number of hours on average people are leaving their homes per day and at what percentage – the blue shows 95th percentile of people are leaving their home for between 16-22 hours per day, while the 80th percentile shows an average ranging from 5-13 hours outside of their home. This is important when it comes to super-spreader activity, social spreading, and number of places people are traveling to and from during the day. 

Algorithms traditionally used by brands to strengthen the relationship with their customers have been repurposed to make epidemiological investigations more precise. A surgical approach of pinpoint accuracy was achieved by identifying the number of encounters and duration of detected encounters and mapping proximity data to the virus. By doing this, potential risk analysis can be provided to help drive effective population engagement and protection.

This also helps medical teams prevent large medical personnel being infected and having to quarantine, taking them away from the healthcare system where they are needed the most. By using specific identifiers, real-time alerts to workers in proximity to possible infections, can be used as front line protection.

Municipalities have a responsibility to monitor and keep their residents safe in line with COVID-19 guidelines while still maintaining essential services and support for the city. Neura’s social distancing index achieves this by leveraging anonymized and aggregated data. By utilizing anonymized data points, Neura is able to build a behavioral model of population density and travel patterns, which subsequently provides cities with predictors and warning flags for future COVID-19 high transmission risk areas. Local governments play an essential role in protecting the population by putting vital measures in place to prevent further outbreaks of the pandemic while safeguarding the economy. 

By generating dashboards with behavioral models that show changes in human interactions and patterns, Neura was able to help the municipality and the city as a whole to protect its citizens.

Encounter Rate Over Time

Action

Neura achieved 70% better efficiency in first responder resource allocation and a 270% decrease in positive cases per day. We were able to do this by providing actionable, real-time, data-driven recommendations to limit individual exposure, inform authorities where additional resources were needed, identify and educate specific populations that weren’t adhering to coronavirus safety regulations, and provide additional resources to return from abroad travelers. 

Key recommendations included splitting up school drop offs, bringing non-essential deliveries to residents, providing information on where extra screening and testing was needed, as well as monitoring adherence to return from abroad travel regulations.

Action 1: Split Up School Drop Off

The first thing the municipality did in response to Neura’s insights to mitigate risk was split up school drop off times. This way crowd gathering was starkly lowered and there was less exposure amongst crowds. Kindergartens were also advised to open multiple gates instead of relying on a single entrance. The data showed that it would be prudent not to operate busses and carpools because these produce virus hotspots, and that testing should be done everyday. 

Non-Essential Delivery:

Next, the municipality workers brought non-essential deliveries to residents in self-isolation, return from abroad travelers, and exclusivity work from home toys for their children, and other similar items to decrease their need to leave the house.

Extra Private Testing:

Further, the municipality commissioned additional private testing for areas Neura identified as high risk for coronavirus, and high transmission. These extra testing centers helped those who were asymptomatic for COVID-19 or who wouldn’t otherwise be able to get tested to ensure they had the additional resources needed to truly impact transmission.  

The next step was to institute solutions fueled by Neura’s data in screening, communication, and public safety. Neura met the authorities in the city at the peak of the second wave at one of the city’s daily meetings that included participation by a full-time operational staff. They discussed everything from food to toy delivery. 

Neura came up with a set of insights showing that there was a transmission rate increase in the industrial zone, more people were clustered inside small areas, meaning information and campaigns needed to be sent about wearing masks in those specific areas. Data collected showed many blind spots, and the following action was taken for each. 

Overseas Return Difficulty:

According to Neura, “We recognized that there were a lot of people coming back from abroad in certain high-risk neighborhoods and suggested running a Facebook campaign encouraging people to complete the necessary isolation period.” Neura also suggested that the municipality obtain a list of the people who came back from the health department to make sure that they are contacted and offered help to help facilitate adherence to the mandatory isolation period. 

The image below shows Neura’s dashboard, sharing which country return from abroad arrivals are coming back from, which helps estimate their probability for coronavirus infection based on the risk of the particular country.

Neura Suggestions:

  • There was a 30% daily increase of people arriving from abroad – specifically into four neighborhoods of the city and there was a high rate of people arriving home from abroad after the summer holidays. A stay at home campaign was launched towards the end of the summer, targeting people who traveled from abroad.
  • There was an increase in non-adherence, which moved from 25% to 30% and then to 60% with non-adherence during busy holidays. There was an urgent need to increase awareness to stay at home. 

The highest risk area was identified and the city directly contacted people who do private screening tests, making use of multiple service providers who do quick tests. They had 300 units to deploy in the area identified as a high risk. The city used private tests, not just state-provided testing because they saw it as their responsibility to provide optimal resources to the most at risk, and in need, areas 

  • People were notified via SMS that there was quick testing being done in their area.

Focus on Communication and Education

Neura suggested where to focus their education efforts, specifically the industrial zone, as a main focal point, and gave them places to focus enforcement efforts in the city during specific times of the day and not just physically crowded spaces, which is relative to the space in question. Based on Neura’s density map, looking at that same day a week before (including specific times), suggestions were made as to where they should put their resources. These included:

  • The industrial zone in the morning time between 7:00am and 10:00am
  • The shopping center in the industrial zone between 10:00am and 12:00pm
  • The shopping center in the city center between 12:00pm and 3:00pm

As demonstrated by using behavior intelligence and acting proactively, Neura’s insights dramatically changed the safety status of the municipality we worked with moving it from a red to a green city in three weeks. 

Equip with the tools Neura provided to identify: crowd gathering, density mapping, social-distancing adherence, return from abroad populations, and virus movement predictions, along with actionable recommendations regarding where to deploy additional resources, testing and screening needs, adherence to return from abroad regulations, and information for where to deploy educational campaigns, this municipality was 70% more efficient in first responder resource allocation and there was a 270% decrease in positive cases per day.

Educational campaigns were created for COVID-19 awareness safety, private testing was brought in for Neura identified high-risk neighborhoods, crowd gathering was monitored and mitigated.  

By fusing behavioral intelligence with epidemiological indicators, Neura’s COVID-19 Emergency Response System is able to map virus movement, identify high risk areas, tell municipalities where coronavirus is transmitted, uncover areas that need additional resources, pinpoint key faults in regulations, and offer amendments. With this information, Neura’s platform can save a city from economic catastrophe, while ensuring the safety of its citizens. 

 

Neura merges with Otonomo bringing mobility intelligence to the mobility, transportation and automotive industries. Read More

×