PandeMe: AI-based Big Data & NLP response against pandemics
Short solution summary:
A solution bespoke for the government and personal uses to automatically extract query-based information from sources around the globe. An AI-assisted platform with advanced automated Topic Modelling, Sentiment Analysis, and Data Structuring; the platform is robust for providing insights by monitoring, analyzing, and comparing global information.
In what city, town, or region is your solution team based?Dhaka, Bangladesh
Who is the Team Lead for your solution?
Mr. Tanvir Hassan Sourov is the team lead, lead for the solution, and the CEO of our company, Socian Ltd. As our Co-Lead and in charge of Operation is Mr. Adib Ahnaf.
Which Challenge Area does your solution most closely address?Respond (Decrease transmission & spread), such as: Optimal preventive interventions & uptake maximization, Cutting through “infodemic” & enabling better response, Data-driven learnings for increased efficacy of interventions
What specific problem are you solving?
We aim to tackle the ineffective, slow, and manual conventional insights for pandemic response with a precise, fast, and automated one enclosing the gaps of disconnected datasets, sector barriers, lack of text AI interpretation of sources, and manual query in information booths. The problems lie mostly in the following three domains:
- Decentralized Dataset:
Datasets of multiple sectors like flight schedules and trade routes are mostly not co-integrated causing a massive ineffectiveness in understanding the macro data-driven picture. Our World in Data has over 18,000 different datasets set for governments to separately analyze.
- Lack of Text AI Adaption:
Citizen-based data-centered pandemic awareness is hindered by the complexity of data while there have not been steps on fastening the process by text AI as much. The significance of these NLP technologies mixed with big data is separately mentioned at https://reliefweb.int/report/w...
- The convenience of traditional data:
While very convenient, traditional data input and visualization are very risky in a pandemic as opposed to big data analysis and NLP.
These are the gaps our solution fulfills robustly.
Who does your solution serve, and what needs of theirs does it address?
Our solution serves the global population, different organizations, and government bodies. Government bodies most affiliated with the solution include ministries like health, finance, and international relations.
We can serve anyone with internet access; even having the ability to integrate a non-smartphone version. The government can also take initiatives to provide information booths to reach individuals with no personal access to the internet. The algorithm can adapt to 'intranet' making it robust even for places without internet penetration.
Our first awareness of the problem came from massive disruption in the economy and COVID-19 maintenance in Bangladesh due to lack of proper big data structure. We've virtually conducted interviews with 17 ministries of Bangladesh and 2 ministry in India to validate the demands. We've shared our views and got extremely positive feedbacks from University of Dhaka and BUET. We went into competitor analysis and most solutions promising to solve the problem ( preferred not to mention) was using decentralized data sets. We were introspective and were inspired by articles like https://www.frontiersin.org/ar... The plan is verified with detailed customer interviews in lean adaption and scraping method of 18,000 consumers and 19 ministry. We can confidently assure our final product to be an ultimate fit.
What is your solution’s stage of development?Pilot: A project, initiative, venture, or organisation deploying its research, product, service, or business/policy model in at least one context or community
Please select all the technologies currently used in your solution:
What “public good” does your solution provide?
One of the direct outcomes of the pandemic is the growing indecision and confusion among the public. With the virus mutating very rapidly and bringing in new pieces of information from different and unreliable sources now and then, a rise of fear among the general public is creating more damage than the virus itself.
Our technology collects voluble data from all possibly relevant data sources and creates a centralized information hub and extracts insights to help the public keep up with the ever-changing trends of the pandemic and reach a decision in their day to day life. From asking about the nearby ICU facilities to staying up-to-date with the latest news on the reopening/ closure of institutions', exams and administrative decisions like lockdown timing, relief goods' availability, our technology will help them tackle this pandemic with proper knowledge and awareness about what's happening where and help them conclude what to do.
This awareness will help the public take data-driven decisions instead of relying on hearsays. As a result, the public will get a sigh of relief from all the chaos and confusion the pandemic has brought.
How will your solution create tangible impact, and for whom?
Healthy general citizens:
The necessity of knowing the spread conditions and heatmap demographics cannot be undermined for general citizen who aim to stay so. Information like availability of ICU, most safe grocery store to current stock pricings can be game changers in keeping them stable. Timely disbursement of accurate data is the prime life saving step.
Tips, advice, protecting against rumor and treatment facilities are some of the major insights used by the infected.
From spread friendly transport restriction to finding the optimal times for lockdown, it is necessary for all insights to be properly analyzed and represented before policy makers. Evidence based policy making could accelerate the recovery from pandemic.
Financial and health based insights as provided via wide arrays of data can help SMEs survive and strategize by being certain amidst uncertainties.
How will you scale your impact over the next one year and the next three years?
Our solutions are auto-scalable and come with load balancers so that if any sudden spike of users is there or there is a huge spike of data on any particular day, the system can auto scale and serve the customers. In the next 1 year, our target is to impact directly or indirectly 10 Million people in the next 1 year and 50 Million people in the next 3 years. And we want to increase our reach over the period of time.
How are you measuring success against your impact goals?
Metrics we are using:
- Net Insights
- Insights per database
- Successive ROI
- Customer Satisfaction
- Policy Impact Comparison (Spread and ROI)
- Net Insights: 1.8M
- Insights per database: 60
- ROI: 160%
- Customer Satisfaction: 100%
- Policy Impact Comparison: N/A
In which countries do you currently operate?
In which countries do you plan to deploy your solution within the next 3 years?
What barriers currently exist for you to accomplish your goals in the next year and the next 3 years? How do you plan to overcome these barriers?
Financial barrier is the number one obstacle towards growing a feasible solution not because of the sustainability model but because of the initial volume fallacy. The minimal data required for such solution is 1 billion blocks. Our annual continuously incoming data requires over 20,000$ worth of server space.
Provided the data needs to be constant, it is necessary for such solution not to only be sold but be sold at a good frequency, at least with targeted subscription. A huge barrier is the time often taken by policy maker to adapt and accept even a proven necessary technology.
While we worked on 40 languages and 100+ dialects; we require certain help in the ones we could not cover yet.
How to overcome them:
For now we have sufficient server backup but will be in need of cloud services along the line. The competition can be a huge connecting point towards multiple sales at due frequency. We are also delegating the tasks of language data collection on the go. There is no knowledge barrier provided our simplistic representation of complex data.
What type of organisation is your solution team?For-profit, including B-Corp or similar models
List any organisations that you are formally affiliated with or working for
- TRR System
- Ping Fei Tech
Why are you applying to The Trinity Challenge?
This pandemic has affected developing countries not just with the virus infections but also with the by-product it has brought. Most of the decisions taken by our policymakers and civil servants are based on assumptions and lack backing by any sort of data.
And the public has become confused and indecisive with the overwhelming amount of information from unreliable sources.
One of the key incidences that kindled this idea in us was when a major Garment Factory decided to restart its activities in the middle of a lockdown which resulted in people rushing into the industrial area which soared the covid 19 cases. Such incoherence among government and private bodies is due to a lack of a robust centralized information hub.
With our technology, we will be able to extract insights that will help big government policies to individual's day-to-day problems. Some of the members of The Trinity Challenge are already working with data extractions and provide data-driven services and by connecting to them, we will be able to maximize and accelerate our endeavour to tackle the growing fear and confusion the pandemic has brought.
What organisations would you like to partner with, why, and how would you like to partner with them?
GSK: Data about sars cov mutations, developments and researches
Internews: With journalists around the world, it provides critical information during the covid crisis
Facebook: Provides data.
Google: Cloud Services
Microsoft: Cloud Services
Palantir: Softwares for data integration
Cubiq: Data and tools provider
Zenysis: provides data integration software to governments
Patrick J McGovern Foundation: AI and Tech
HKU med: Data about sars cov mutations, developments and researches
Imperial College London: Data about sars cov mutations, developments and researches