by Davinder Kohli

Over the last decade, online retail companies have successfully, metaphorically speaking, constructed the DNA of their customers. As a result, eCommerce companies can now precisely market specific products to customers based on their purchase history or browsing patterns. Companies delight their customers by enabling them to purchase products with a single click, providing both price and shipment tracking transparency. Contrarily, the healthcare industry significantly trails behind the retail industry in providing such an engaging experience to its consumers and thus giving rise to healthcare consumerism. Healthcare consumerism is designed to enable patients to become more fully involved in their healthcare decisions. It helps transform a patient’s health benefit plan, putting decision-making and economic purchasing power in the hands of plan participants.

Patients Concerns

Healthcare consumers are concerned about their long-term care expenses, chronic care, and end-of-life care. This need is driving customers to better understand their benefits, treatment coverage, and medical costs in an easily consumable manner. Furthermore, healthcare customers demand a better experience when shopping for a doctor, ordering refills, and monitoring their health metrics. Healthcare providers and payors can certainly learn a lot from the tactics employed by the retail industry to construct a needs-based profile of their healthcare consumers to effectively engage them. Constructing and analyzing the “customer journey map” of a healthcare consumer reveals both the potential opportunities and pitfalls of consumer touchpoints with the healthcare system (providers and payors). Broadly speaking, there are three phases of the healthcare consumer’s interaction with the healthcare system: 1) Before Treatment, 2) During Treatment, and 3) Post Treatment. The figure below shows the activities during each of these phases.

Monitoring Chronic Care Patients

A common observation across all of these phases is the use of wearables such as Fitbit and remote patient monitoring (RPM) devices such as pulse oximeters, glucometers, and blood pressure monitors. Although RPM devices have been around for the last 20 years, COVID-19 has accelerated their proliferation along with telemedicine, specifically for monitoring chronic care patients. These wearables and RPM devices can help providers and payors construct a profile of their healthcare consumers. The data transmitted from these RPM devices are monitored for any irregularities in patients’ biometrics by Care Managers, usually hired by providers to provide extended care. Should there be any fluctuations in a patient’s predetermined biometric threshold, Care Managers reach out to the chronic care patient advising them to adjust their medication level or schedule a TeleVisit with their primary care physician (PCP).

While this is an extremely useful service that meets the promise of value-based care, it is neither economically scalable nor can forecast fluctuations in patients’ biometrics. Moreover, Care Managers cannot discover patterns, such as causes of irregularities in patient biometrics, across the enormous amounts of the data collected from these RPM devices. This is where an artificial intelligence (AI)-based Care Manager (AICM) system can help.

AI-based Care Manager

An AICM is not intended to replace a Care Manager but help achieve economies of scale in chronic care patient monitoring and to predict irregular conditions before they occur. Along with making predictions about chronic care patients’ health and timely alerting the patient, the AICM can automate the scheduling of patients’ TeleVisits, order medication refills, answer patient’s dietary questions all based on analyzing patients’ real-time condition. As AICM continues to learn about the patient’s daily patterns from wearables and RPM device transmitted data, it will establish patient’s 360. Healthcare companies can depend on AICM to analyze patient habits and diets that influence their biometrics to better devise personalized treatment plans. On the other hand, AICM can be used by healthcare consumers to get recommendations, based on their health condition, the changes to their diet and lifestyle for their well-being.

Empowering patients, providing personal care, and leveraging data and intelligent technologies now serve as the backbone for optimizing outcomes. The use of artificial intelligence is just beginning to alter the way clinical providers make pertinent decisions regarding patient care and healthcare operations.

Doing a gut check on AI

by Davinder Kohli

Growing up in an Indian household, I commonly heard that for a body to be healthy, its gut needs to be strong. To accomplish this, the emphasis was given to the types of food to consume and when they were to be eaten. For example, when preparing dishes such as spinach or cauliflower, garlic and ginger were added not only for enhancing the taste but for curbing any digestive flatulence. We were also specifically told not to consume fish with milk and to keep from eating watermelon at dinner.

While this advice is still prevalent and might have served me well, there is nothing scientific about it. Everyone’s body reacts differently to the same food and produces different glycemic levels depending on the microbiome in their gut. Microbiome is an ecosystem of trillions of microorganisms (bacteria, fungi, viruses) that live in our bodies. According to medical journals, the microbiome is even labeled ‘a supporting organ’ because it plays so many key roles in promoting the smooth daily operation of the human body.

In my recent webinar, I was fortunate to be in a virtual fireside chat with Naveen Jain, CEO of Viome, Inc., a precision health and systems biology company that offers services and health tests to help people understand their unique biology in areas like gut microbiome, cellular, and mitochondrial health. Analyzing the microbiome to understand the effect of foods that we consume is something the company does using artificial intelligence (AI) technology.

The focus of this talk was on what it takes to launch an AI initiative and the multidimensional challenges involved in bringing AI-based solutions to the market. Additionally, we discussed the risks of having autonomous systems with an unsupervised learning algorithm, finding the right talent, and the future of work. For those who cannot spend an hour going through the video, below are some of the key takeaways of my conversation with Jain.

–Data collection and modeling in AI

Collecting reliable data and its importance in constructing the right model with relevant parameters is an immense challenge in building AI systems. Jain advises checking the source, quality, and relevance of publicly available datasets before using that data for training AI models.

A better approach for organizations is to conduct their own survey, collect metadata and data, and test which of the collected metadata is applicable in the model. The data should be collected incrementally with a certain sampling size and then processed through the model. A feedback loop must exist in the AI system that enables data scientists to continually analyze the output of the model and accordingly adjust the input metadata, training datasets, and algorithms.

–Noise in data

There is a direct correlation between the number of features and the amount of homogeneous data to be collected. The higher the number of metadata elements in an AI model, the more data required. For example, if an AI model analyzes 1,000 features, collecting only 100 data records will not yield the right outcome. Jain suggests reducing the number of features if large amount of homogeneous data cannot be collected.

Additionally, collecting large amounts of data also reduces noise (use of irrelevant parameters). The randomness of certain parameters may seem large in a small dataset but might diminish when analyzing large datasets. Contrarily, what may seem like a random parameter might get amplified if that parameter seems to be persistent across larger datasets. Another noteworthy point is to use several models to validate the outcome rather than relying on the outcome of a singular model.

–Data biases and data privacy

Data models and algorithms are designed by fallible human beings. Of course, most of these systems are coded with the best intentions. Yet, intentionally or inadvertently, biases do exist in data, models, and algorithms. Jain advises against using a single person to code the algorithm or model. Models used by AI should not be presented as black boxes to the consumers, rather explained in terms of the parameters considered in producing the outcome.

When collecting personal data from consumers, either personally identifiable information (PII) or personal health information (PHI), the data should be encrypted and stored separately from the rest of the data to ensure Health Insurance Portability and Accountability Act or HIPAA compliance. However, a key differentiator for Viome, when it comes to data privacy, is that the company stores a person’s gene expression which is constantly changing. Therefore, getting access to that data would be futile for anyone outside the organization.

–The future of work

It is commonly debated that AI systems will eliminate many repetitive jobs in the future. While it is true that industrial revolutions result in manual jobs being taken over by automation, they create many more new jobs thereby producing higher throughput and enhanced productivity. Those who feel threatened by AI taking away their job should see it as an innovation that removes repetitive tasks and enables them to focus on the human experience.

On the flip side, finding talent to build AI systems is a challenge. Talented workers such as mathematicians, programmers, and data scientists are required. When I asked Jain how we can overcome this barrier, his response was that talent finds those who are solving big problems for human mankind. People want to work for companies that are audacious and solve challenging issues.

Bottom line: I hope you gained some new knowledge from my conversation with Naveen Jain, CEO of Viome. In many ways, AI is still a nascent technology that will eventually change the world. In fact, when I asked Jain about the future of AI, he was quick to respond with a witty comment that AI will become an acronym for Ancient Intelligence, and artificial intelligence will be the “new intelligence.”

by Norman Murrin

The best ways to achieve certification, benefits of certification and how to motivate yourself.

One of Creospan’s client is well underway moving to AWS and I have achieved the three AWS Associate level certifications. Our client would love to see more of their employees investing in themselves. However, most developers don’t make the time for certifications or other professional improvement outside of work because they don’t understand how important it is for their career.

Most developers believe that learning on the job is the most effective way to improve. There are indeed many benefits for you, the developer, to learn on the job, but not many benefits for your employer or client. Or rather, your client would benefit much more from you leveling up on your own time and bringing new knowledge to work. You will benefit in new ways too when your clients recognize that you are operating on a different level of professionalism.

There are plenty of effective ways to learn a new technology, and I do not want to convey that certifications are better than taking an online course, creating a sample application or attending a conference. All are effective, but certifications have some distinct benefits that work well for me.

Broad Spectrum Knowledge

Most certifications are put together in a way that requires the candidate to understand the complete breadth of a technology or platform, but not necessarily to understand every nuance or detail. Engineers holding certifications should be able to participate in design discussions and effectively complete high level design work.

Many will expect that a certification holder knows all the details and can develop any part of the solution effectively. However, that is generally not true, nor is it generally the point of a certification. Understanding a technology broadly is the point, and the in depth knowledge is learned through experience and referencing available documentation.

Your client will benefit more by you understanding the technology broadly then by your having some in depth knowledge on parts of the technology. Experienced developers can figure out the details given a task. However, getting the broad strokes of a project wrong can be very costly.

Motivation and a Goal Built In

One aspect of certifications that works very well for me is that the motivation and an end goal are built in. I generally will make an initial investment in a book or online course. This initial investment provides a bit of motivation to continue. Next I put in some time, my own time, providing more motivation to continue.

When I feel that I am making progress and starting to gain an understanding, then I will often sign up for the certification exam. This generally costs upwards of $100. More importantly, it sets a deadline, usually another week or two out. This time is used to buckle down and make sure I have done all that I can to pass the exam.

Proof of Achievement

Once you have achieved certification then you can make sure others know it too by putting it on your resume, attaching to your social media, and even putting it in your email signature if the certification is core to your career. It is this aspect of certification that can make it a vehicle for venturing into new areas. Often clients will have you on a project when you lack experience, but happen to have a certification. As a team leader, it is often very helpful to have consultants set certification goals because it is measurable.

It’s Your Career

You have chosen to work in technology — choose to be a professional. Put in the time to be effective for your employer or client and to be as valuable as you can to your team. Certifications can be a valuable tool on your path.

At Creospan, we are always looking for consultants that realize their responsibility to improve professionally and we make training readily available. Of course this is necessary to get through client interviews and perform to client expectations. However, it’s always helpful when consultants achieve a certification. Clients value certifications and often will skip that portion of the interview. Consultants achieve a level of confidence with certification that shows at the client.

by Joseph Socarras

Recently one of our clients, a startup video content sponsorship platform, approached us at Creospan after deciding to leverage AI to innovate their existing business model. Currently, our client provides sponsorship opportunities on content before that content is posted anywhere. Our client determined that if they could figure out a way to predict user engagement on their content before posting that content on social media, then they could sell sponsorship deals much more effectively. They decided on using AI to predict content engagement, reasoning that AI trained on vast amounts of historical data can sometimes outperform even the predictions of human experts.
After discussing the problem statement and reviewing available historic data with our client, we decided we’d need about four weeks to build a starter AI model. We began by exploring what data was available to us and which AI models we should try implementing. Since our client had access to an extensive database of video metadata and engagement rates, we reviewed and cleansed that available data. Using a combination of Python and Pandas, we removed missing or incomplete data points, standardized data fields such as dates, derived new fields like content age and content engagement rates, and prepped the data to be consumed by an AI model. 
The next challenge we faced was which AI modeling approach to select. Since we were attempting to predict what specific level of engagement a post would receive, we decided to implement an AI model known as a regression model. Next, we decided which regression models we’d try based on the amount and type of data we had accessible to us. In this case, we avoided data-hungry models such as neural networks, and instead explored other modeling approaches including Random Forests and linear & logistic regression models. After evaluating preliminary performance, the client and we selected a lasso regression modeling approach. Once we settled on an AI technique, we built an initial model, then deployed the model as an API that our client can consume whenever they like. Once done, Creospan handed off further improvement and maintenance of the starter AI model to our client. 
Here are some examples of predictions made so far:

Although predictive analytics challenges can seem daunting at first, they can be addressed methodically. In the case described above, the Creospan team worked with our client to acquire, process, and analyze large amounts of relevant social media data. Afterward, we used that data to train and test various regression models to predict user engagement, selected the best performing model, then operationalized that AI model as an API for usage anytime within the client’s website & mobile app. Here is a visual summary of the process we followed:

As a result of Creospan’s efforts, our client is now able to automatically predict their content’s expected engagement rate (to a reasonable degree of accuracy), allowing them to sell content sponsorship deals much more efficiently and effectively. By now you have hopefully learned that AI can generate strong, unique business value, and that it doesn’t have to be hard for your organization to start leveraging AI today! 

COVID-19 has caused health crises leading to financial crises across several industries. Companies are struggling to cope with the crises and looking for ways to continue to engage with their customers in a manner that is not only safe and delightful but can potentially open up new revenue streams. More than ever, its imperative for organizations to innovate and transform digitally. This does not mean that we throw more technology at problems but this requires a concerted effort within an organization, both at the cultural and process levels. In this webinar, we talk to two technology executives from the healthcare and energy industries on determining how they are achieving digital dexterity in today’s climate.

by Scott Kramer

You have probably noticed that the market for much of the Tech workers has collapsed faster than IBM’s BlueMix brand.  During the current recession, bordering on depression, we are offered 2 worlds, one world where it seems scary but the talking heads tell us it’s ‘fine’, we are in a ‘V’ shape recovery and companies are making more money than ever – examples like Amazon, Netflix, Zoom, Grainger, Chewy, GrubHub, etc.  It’s a world where the Federal Government is printing money faster than ever before (thus hiding taxing you faster than ever before) and giving money away for free.  The Second world, the one that all of us want to ignore, is where the danger lurks.  This world features many industries that are crushed by Corona like Hyatt, Walgreens, Wells Fargo, AT&T, Mercedes, United, Lyft, etc.  This world has significant drops in the number of projects (thus jobs), rampant unemployment, and extreme uncertainty.  

Both of these worlds are one and the same and many companies are somewhere in the middle of the two scenarios. In either case if half of the tech industry is depressed, it has a huge effect on the tech market. The helicopter drop of free money is helping us avoid an outright depression, but the drop in projects and jobs is crushing compensation (some estimates are that it dropped 50% at one point).  But what does the future hold?  Until sometime in 2021, the tech world will continue on its ‘slow-roll’, some benefiting, others getting by, and some losing.  In general no one wants to risk investing with so much uncertainty.  So don’t expect a V shaped rebound, but do expect a slight bump as we near the 4th quarter.  Big companies must spend unused budget money or risk losing that budget money next year.  

A reasonable assumption is that sometime in 2021, the full tech market we all love will start to go back to normal.  Large bumps in contractor projects will signal a market reboot as companies need to jump start projects and are still risk adverse to full time hires.  With a reduction in uncertainty, those back-burnered critical projects will become relevant again.  In the meantime, companies will continue to scale back projects, lowball compensation, and it’s going to be a really tough period.  The best advice is to get any job or keep your existing job until 2021 and as soon as the storm passes reevaluate your position.

Sources:

https://www.ibm.com/cloud/blog/announcements/bluemix-is-now-ibm-cloud

US Industries Being Devastated by the Coronavirus

https://www.grunge.com/195110/companies-that-are-actually-benefiting-from-the-coronavirus/

By Edward Tanko

In this article I am providing a quick start guide to docker swarm commands by working through a three nodes docker swarm cluster.

Please note, you should already be familiar with docker before going through this guide.

Definitions

• Docker swarm = Docker in swarm mode 
• A swarm consists of one or more nodes running Docker Engine 1.12 or later in swarm mode.
• Each node in the cluster can be a manager or a worker or both
• A node is an instance of the Docker engine participating in the swarm
• To deploy your application to a swarm, you submit a service definition to a manager node. The manager node dispatches units of work called tasks to worker nodes.
• Manager nodes also perform the orchestration and cluster management functions required to maintain the desired state of the swarm. Manager nodes elect a single leader to conduct orchestration tasks.

Setup Docker cluster

• Make sure you have three Servers (Linux VPS) running over your network or the Internet with each node having a dedicated IP address. This machines will be the nodes on the docker clusters
• I recommend you create an Amazon EC2 instances or a Digital Ocean Droplet
• Install Docker in each of the node (easy installing command curl -sSL https://get.docker.com | sh). This command will pull a docker installation script and run it on the node.
• In this tutorials, we will refer to the nodes as (Node-01, Node-02 and Node-03)
• On Node-01, run the command

docker swarm init --advertise-addr <Node-01 IPAddress>

This command will set up the docker swarm and enable Node-01 as the manager node on the swarm.

• The next step will be to add Node-02 and Node-03 as nodes on the swarm.
  • To do this run this command on Manager Node, Node-01

docker swarm join-token manager # For a manager node token
docker swarm join-token worker # For a worker node token

This command will display the docker swarm join token which is needed to add other nodes to the cluster or swarm as a manager or worker node. It will look like this

docker swarm join --token <swarm token> <Node-01 IP>:2377

Copy the join cluster command and run on the two other nodes to add them to the cluster. This will create a three node cluster with Node-01 as the lead manager node.

• Another way to upgrade a worker node (e.g Node-02) to a manager node, will be to join as a worker and then run this command on Node-01

docker node update --role manager Node-02

Only manager nodes have access to the swarm commands, the worker nodes are just docker environments to execute tasks.

• To see that cluster created, run on the manager node

docker node ls

Creating a Service

• Service is the tasks to execute on the nodes.
• When you create a service, you specify which container image to use and which commands to execute inside running containers.

docker service create --name <ServiceName> -p <exposePort>:<serviceTaskPort> <image_name>:<image_tag>

• To create 3 replicas service use

docker service create --name <ServiceName> --replicas 3 -p <exposePort>:<serviceTaskPort> <image_name>:<image_tag>

• To see all the services created use

docker service ls

Network

Overlay network

Creates an internal private network that spans across all the nodes participating in the swarm cluster. So, Overlay networks facilitate communication between a docker swarm service and a standalone container, or between two standalone containers on different Docker Daemons.

• Use this command to create an overlay network on your nodes

docker network create --driver overlay <network_name>

• To see all network created do

docker network ls

• To create a service in a network use

docker service create --name <ServiceName> --network <network_name> -p <exposePort>:<serviceTaskPort> <image_name>:<image_tag>

Docker Stacks

Stacks accepts compose files as their declarative definition for services, network and volumes. docker stack deploy

• To deploy your application as a docker stack, define your application in a yml compose file. the run the command

docker stack deploy -c <compose-file.yml> <stack_name>

• To see all running stacks use

docker stack ls

• To list all task in the stack use

docker stack ps <stack_name>

• To remove stack use

docker stack rm <stack_name>

• To list all services in a stack you can use

docker stack services <stack_name>

• Example of a stack compose file

version: "3.8"
services:
  db:
    image: postgres:9.4
    volumes:
      - db-data:/var/lib/postgresql/data
    networks:
      - backend
    deploy:
      placement:
        constraints: [node.role == manager]

-   The Example will create a service call db using the image postgres:9.4 and run on a network call `backend`. 
-   The constraint property on the deploy tell the swarn that This service can only be deployed on a node with a manager role

Docker Secret

• Only stored on Disk on Manager nodes
• Secret are assign to a service
• They looks like files in container but are actually in-memory
  • /run/secrets/<secret_name> or /run/secrets/<secret_alias> To create a secret

docker secret create <secret_name> <secret.txt> # using a file

OR

docker secret create <secret_name> - # Then type the screat in std input

• Available commands

  create      Create a secret from a file or STDIN as content
  inspect     Display detailed information on one or more secrets
  ls          List secrets
  rm          Remove one or more secrets

• When a secret is create, you will never be able to see it value. you can only assign it to a service
• Ex: let’s create psql service and pass it password

docker service create --name psql --secret psql_user --secret psql_pwd -e POSTGRES_PASSWORD_FILE=/run/secrets/psql_pwd -e POSTGRES_USER_FILE=/run/secrets/psql_user postgres:9.4

• Using a stack compose file: We can pre-create the secret or we can add the secrete in our compose file

1. Pre-create: This means that psql_user and psql_pwd would have already been created in the swarm using docker secret create

file: docker-compose.yml

version: "3.8"
services:
  psql:
    image: postgres:9.4
    secrets:
        - psql_user
        - psql_pwd
    environment:
        POSTGRES_USER_FILE=/run/secrets/psql_user
        POSTGRES_PASSWORD_FILE=/run/secrets/psql_pwd
    volumes:
      - psql-data:/var/lib/postgresql/data
  
secrets:
    psql_user:
        external: true
    psql_pwd:
        external: true

volumes:
    psql-data:
    

1. Creating the the secrete in the stack compose file on the fly

file: docker-compose.yml

version: "3.8"
services:
  psql:
    image: postgres:9.4
    secrets:
        - psql_user
        - psql_pwd
    environment:
        POSTGRES_USER_FILE: /run/secrets/psql_user
        POSTGRES_PASSWORD_FILE: /run/secrets/psql_pwd
    volumes:
      - psql-data:/var/lib/postgresql/data

secret:
    psql_user:
        file: ./psql_user.txt
    psql_pwd:
        file: ./psql_pwd.txt
volumes:
    psql-data:

• Make sure the compose file and the psql_user.txt and psql_pwd.txt are in same directory when creating the stack
• with this format, if we remove our stack, it will also remove all the secrets created

Docker healthchecks

• Support in Dockerfile, Compose yml, docker run, Swarm service
• Docker engine will exec the command in the container
  • e.g curl localhost
• It expect exit 0 (OK) or exit 1 (Error) we could also have false for error
• Three container state: starting, healthy, unhealthy
• We can see the status of our container when we do
  • docker container ls or
  • docker container inspect : check last 5 healthcheck
• Docker run do not take action when the healthcheck is fails. Just needed for status check
• In swarm, Healthcheck is important!!!
  • Services will replace tasks if they fail healthcheck
• Ex: Healthcheck in Docker run

 docker run \
   -- health-cmd "curl -f localhost/_cluster/health || false" \
   -- health-interval=5s \
   -- health-retries=3s
   -- health-timeout=2s
   -- health-start-period=15s \
 elasticsearch:2

• Health check in a compose stack file

version "3.8"
services:
    web:
      image: eddytnk/myapp:latest
      healthcheck:
          test: ["CMD", "curl", "-f", "http://localhost"]
          interval: 1m30s
          timeout: 10s
          retries: 3
          start_period: 40s

by Davinder Kohli

The “Future of Work” was a popular topic in the conversations and presentations I attended at the World Economic Forum’s Annual Meeting of the New Champions held in Tianjin, China in 2018. That year, for the first time, the Annual Meeting brought together experts and decision-makers in the World Economic Forum’s newly launched Fourth Industrial Revolution Councils. The mission of these councils is to help policymakers and businesses strike the right balance between enabling emerging technologies and proactively mitigating the social risks that can result.

The Fourth Industrial Revolution, categorized by the fusion of technologies, involved the combination and transformation of different core technologies to create new products and markets. These included Artificial Intelligence (AI), the Internet of Things, Blockchain, autonomous and urban mobility, drones, and precision medicine.

According to Dr. Klaus Schwab, founder and executive chairman of the World Economic Forum, compared to the previous industrial revolutions, the Fourth Industrial Revolution is evolving at an exponential rather than a linear pace. While all this rapid innovation and transformation has the power to improve the quality of life for populations around the world, this revolution could yield greater economic inequality.

Dr. Schwab further discussed how automation will result in a net increase in rewarding jobs. However, the overall job market will be segregated into “low-skill/low-pay” and “high-skill/high-pay” segments, which in turn could lead to an increase in social tensions. Additionally, this “haves-and-have-not” scenario could result in degraded healthcare in lower economic segments of society.

Social Determinants of Health

Little did we imagine in 2018 that a pandemic like COVID-19 would occur and cause a draconian shift in the labor market. Adoption of a low-touch economy and disruptions in the supply chain, initiated by COVID, have pushed businesses to further automate their operations. While this punctuated equilibrium is necessary, we must look out for those that are highly vulnerable to disasters and impacted through this automation.

The level of the population’s vulnerability to disasters is determined by what is known as the Social Vulnerability Index (SVI). SVI is a weighted score of the different elements of the Social Determinants of Health (SDoH). According to the World Health Organization, SDoH is the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life. COVID has shifted the tectonic plates of SDoH and its imperative for healthcare companies to understand this shift.

Impact on Healthcare

Among the many challenges faced by healthcare companies lowering the overall cost of care and improving patient outcomes is pivotal. A big part of healthcare costs involves treating patients with chronic conditions, emergency room (ER) visits, and readmissions. Some of these costs could be reduced or avoided by increasing patient engagement, especially those with a higher SVI.

The key is to procure and link different data elements identifying the socioeconomic (education, poverty, income, employment), housing and transportation (housing density, mobile homes, vehicle access), household composition (age over 65, age below 18, disability), and sociological (minority status, language) conditions of the patient population.

By combining this social structure data with the healthcare provider, patient-generated, and financial data using advanced analytics (graph databases and predictive algorithms), healthcare services can be designed and personalized around the patient.

Patient-centric Care

These personalized healthcare services should entail key areas such as providing health and therapeutics education, meal plans, access to transportation, payment services, community support, and active health monitoring. Tech-business leaders should enable these services through cloud and mobile-based digital platforms, providing patients ubiquitous access in a comprehensive manner.

Healthcare firms could partner up with ride-share companies to provide transportation access to patients for regular checkups and early diagnosis, creating a win-win situation for both providers and patients and helping to avoid unnecessary visits to the ER. Similarly, partnering up with food delivery services and enabling patients to seek nutritional or special dietary meals during treatment and post-admission hospitalization can prevent readmission’s.

Additionally, tech-business leaders need to make these digital platforms intelligent using AI technologies, such as natural language processing (NLP). This will increase patient engagement as such platforms can provide contextually relevant information without having the patients wait for getting answers from providers or spending countless hours scouring the internet.

As healthcare companies explore opportunities to become more patient-centric, they need to be mindful of shifts in the labor market and the increase in SVI caused by the Fourth Industrial Revolution and accelerated by the current pandemic. While such awareness will help bring down healthcare costs and increase the quality of patient health, it also has the potential to reduce the “low-skill/low-pay” population’s vulnerability to disasters.

by Davinder Kohli

In 2014, when I submitted my presentation titled, “Transforming Business-driven to Technology-driven Organizations” to an annual technology conference from Oracle, the title received some resistance from some business and technology leaders alike.

Their argument was “How can organizations be led by technology when technology is just an enabler? The technology provides solutions to business problems; therefore, organizations are business-driven.” This argument makes a lot of sense in organizations that think of technology (or IT) simply as their backbone of achieving operational excellence. Solving the business problem of accurately and expeditiously triaging customer contracts using OCR (Optical Character Recognition) technique, is an example of operational excellence.

However, just having operational excellence is not enough to respond to the threats of disruption or delighting customers. Organizations must be technology-driven to recognize disruptions, respond to them, and capture new markets. This requires current technology leaders to be business leaders. They must think about how to grow the company revenue while keeping an eye on their internal technical landscape and strengths.

Recognize Disruptions

Undeniably, we live in a digital economy where everything is hyperconnected and we have the tremendous computing power to process data. AWS, Microsoft Azure, and Google Cloud Platform have leveled the playing field, enabling the smallest of the companies to compete with large organizations. Top management schools provide programs in artificial intelligence (AI), blockchain, and data analytics, to prep the next generation of business leaders on technology.

Therefore, it is imperative for technology leaders to not pride themselves on building moats using technology. The new tech leader must focus on value creation and value realization. Value creation requires looking at how existing business models are getting disrupted in favor of new models that remove existing customer pain points and also attract new customer segments.

For example, the traditional auto insurance industry model of charging a monthly premium to individuals whether they drive regularly or occasionally was disrupted by a company called Milewise, later acquired by Allstate. Milewise from Allstate provides pay-per-mile car insurance catering to infrequent drivers. Technology leaders should be building innovation centers dedicated to experimenting with new business models. They should then demonstrate and test those learnings in the market, alongside startup companies to remain competitive.

Delight Customers

Tech leaders in large organizations must pivot from their current role of just being custodians of IT to proactively understanding their customers to effectively design products and services around them. They must identify ways of using technology to tap into the next generation of customers.

For example, tech leaders may leverage different social media channels, such as Instagram and TikTok, and partner up with influencers in those channels to discover new customer behaviors and trends. Based on these trends, technology leaders must identify solutions that will delight their customers. Companies like Wayfair had started experimenting with emerging technologies such an augmented reality (AR) and now provides a mobile AR app for their consumers to visualize furniture in their homes before purchasing.

Keep in mind, ever-changing and evolving customer behavior has been accelerated by COVID-19. Besides changing customers’ purchasing capacity and channels of distribution, it has accelerated the adoption of the do-it-yourself (DIY) mindset. The new tech leader must account for changing geopolitical, socioeconomic, and diversity landscapes when launching new offerings into the market.

Build Partnerships

As technology leaders become business savvy, the partnership model changes from technology-business to simply a business partnership. The new tech leader has the onus of bringing together the talent across different silos to build integrated technology solutions that enable the organization’s agility to respond to current market trends. Additionally, the leader needs to look externally to establish partnerships with companies that can help bring new capabilities and augment internal talent.

As the new tech leader focuses on achieving operational excellence, responding to disruptions, and building partnerships, the overarching mission should be to enhance the customer value proposition.

by Davinder Kohli

Just before we entered the world crises caused by COVID-19, I was asked what problems we are solving by building expertise in Knowledge Graphs. My answer was simply, “We are sharpening the spear.” However, there is more to it than a simple one-line answer.

Knowledge Graphs technology has been around for over seven years when Google first started using it to enhance the value of the query information returned by the search engine. The technology enables us to seamlessly connect and analyze different pieces of disjointed information that relational database technologies cannot deliver. Knowledge Graphs, and its underlying Graph Database technology, is used by popular social media, movie streaming, and eCommerce companies in recommending new relationships, movies, and other products to their audiences.

As this technology has the potential to be a game-changer in companies’ digital transformation journey, I have been cursorily reading about Knowledge Graphs for some time but never had an opportunity to do a deep dive into it until earlier this year. My team and I decided to start looking into it as part of our organization’s Data and Analytics practice to build machine learning solutions.

Crises and Opportunity

With the current health and financial crises caused by COVID-19, businesses, specifically those that have front-line workers, are struggling to find answers to how COVID has impacted them and how they can move back to growth. To find solutions to these difficult problems, a lot of disjointed but pertinent data needs to be procured, connected, and then analyzed. As the saying goes, “Luck Is what happens when preparation meets opportunity.” In response to the current crises, our team collectively brainstormed and figured out that Knowledge Graphs technology would be one of the pivotal components of our solution platform.

For a business to understand the impact of COVID-19 and its level of readiness to resume operations, a holistic intelligence platform is required. As our team of data scientists continues to analyze multiple dimensions, a few have come to light — taskforce, supply chain, and facility readiness. Each of these dimensions is based on connecting data elements from curated data sources such as COVID data lakes, general population with economic data, and, of course, company-specific employee data. All the curated data comes from trusted sources that have been cleansed, enriched, and wrangled, with personally identifiable information (PII) removed.

Knowledge Graphs vs. Relational Databases

Leveraging Knowledge Graphs to unlock insights from curated datasets requires building data models and implementing algorithms. Data models have traditionally been designed keeping relational database technologies in mind. Insights that need to be gleaned drive the designing of data models; putting functional constraints on their usability. Any change to the desired insights or adding new insights requires changes to the underlying data models. Simple algorithms implemented in relational databases can get overly complex as the number of relations (known as JOINs in tech speak) among different data elements increase. These algorithms when combined with large datasets can cause them to run for days or at times simply collapse due to the CPU and memory limitations of the underlying hardware infrastructure.

This is where Knowledge Graphs shine. Data models in Knowledge Graphs are created based on natural connections between data elements and without requiring upfront knowledge of the insights to be revealed. This fosters the discovery of new patterns and insights without changing any underlying data models.

There are multiple vendors of Graph Database technology. For scalability and speed purposes, we chose native Graph Database technology from TigerGraph. TigerGraph enables executing algorithms with 5+ levels of connectedness (also known as hops) among the data elements in sub-seconds. As the number of hops increases combined with the quality of data, the accuracy of the insights increases.

COVID Related Insights

At Creospan, we continue to enrich our datasets, graph models, and augmenting the relationships among existing and new data elements. Our goal is to unveil as many new insights as possible to help healthcare, manufacturing, retail, and financial companies determine their pandemic response and move-forward strategies. We see Knowledge Graph technology as an integral component in the ecosystem of an organization that is considering to connect disjointed information spread across the enterprise to glean insights for driving better business outcomes.