Inspiration

During the Covid-19 pandemic, we were constantly looking for Covid-19 resources and information regarding policy and government updates, vaccine updates, etc. Updated and reliable data can be hard to find, as information is often scattered across the internet. Additionally, relevant scientific papers and articles are hard to access as well, especially for those who are not associated with an educational institution.

Severe acute respiratory syndrome (SARS) broke out in 2003 and Middle East respiratory syndrome (MERS) broke out in 2012. Covid-19 comes from the same family as SARS and MERS. Based on this trend, there is a possibility that more viruses from this family can infect humans around the world.

A scientific article from Science of the Total Environment suggests that climate change and globalization may be causing factors of SARS, MERS, and Covid-19. With global warming on the rise and air travel becoming more and more convenient, everyone will have to face many more pandemics in the future.

The Covid-19 virus is also constantly evolving and adapting, so we may have to face many Covid-19 variants in the upcoming years.

With the high possibility of more pandemics occurring in the future, SuperFem Gliders created “streaMLined Pandemic Directory", a free web app that doubles as your one stop destination to find reliable pandemic resources. Along with having a centralized database of pandemic resources (verified scientific information about the causes and effects, and what to do in case of a pandemic), we decided to build a machine learning model that could predict the start and outcome of a pandemic, to help the common citizen and the technical scientist have access to such information.

What it does

streaMLined pandemic directory is a free web application designed to be a preparation system for future pandemics.

Our web app has three main features:

  1. A machine learning algorithm that uses imputed data and relevant factors to predict the outbreak changes and values. Scientists can use values such as the number of new cases and deaths in their country, the percentage of population that is vaccinated, and any relevant data into the algorithm. The algorithm will use the data to generate predictions, which can then be used to prepare for future pandemics.

  2. Data visualizations that display predictions made from the machine learning algorithm. Once the algorithm makes predictions, the predictions will be used to create graphs that the general public can understand. This helps make our web application more user friendly for everyone. StreaMLined pandemic directory will be updated with new graphs when new information is imputed into the machine learning algorithm.

  3. A pandemic resources page that includes links to reliable scientific papers concerning pandemics. All the scientific papers present on this page are collected from trusted databases, such as Google Scholar. Governments, scientists, and the general public can access the links to find professional and updated information about current or future pandemics. In the future, we can upgrade the page to feature a verification system that allows other scientists to add scientific articles to the webpage. Scientists can submit a research paper, and the verification system will check the paper for reliability.

Aside from the three features, streaMLined pandemic directory was designed to be used by everyone, from the everyday worker to epidemiologists from around the world. Our web application will help prepare governments and citizens for future pandemics. With all the resources and data gathered in one place, scientists and the general public won’t have trouble finding lots of reliable and updated information.

How we built it

The machine learning algorithm was programmed in Python and is run on PyWebIO for now. The simple data visualization and ML model were created with jupyternotebook. The ML model was deployed using pickle by converting the notebook into a pickle file, then with python and libraries such as flask and PyWebIO. The web application itself was developed on QOOM through frontend code using css, html and vue.js.

Challenges we ran into

  • Using a ML model & designing our webapp according to our plan was quite challenging.
  • It took a lot of time coordinating our team meetings as the time difference in our group is up to 12.5 hours.
  • Through the process of creating streaMLined pandemic directory, we had to adapt and learn to use new platforms, such as PyWebIO, Qoom, and Figma. We ran into several minor problems along the way, but we were able to produce a final project we are proud of.

Accomplishments that we're proud of

  • We were introduced to a new Python library named PyWebIO, which we used to deploy our machine learning model to webapp. We definitely want to work more on PyWebIO as it is a new platform.
  • We created a prototype on Figma with several slides that show how we envision our webapp to look, and how the three features of streaMLined pandemic directory fit together on a website.
  • We created our web application on Qoom, complete with all three features of streaMLined pandemic directory, and a Typeform for people to submit questions/concerns.

What we learned

  • We learnt about the Python library PyWebIO and learnt to successfully operate it such that we could run our machine learning model on it, which resulted in a lot of fun and a sense of accomplishment.
  • We learnt how to better manage our time since this was the shortest hackathon any member of our team had worked in.
  • Even with a major time difference of 12.5 hours, we were still able to collaborate as we were in the same time zone, teaching us that time difference is a small obstacle to tackle in the face of passion.

What's next for streaMLined Pandemic Directory

Once we launch our web application, we plan on developing more advanced machine learning models to make a wider range of predictions. The simple ML web app made with Pywebio is just to show how the ML section should work in the actual app. Therefore, we are going to deploy the ML app to our final product in the future. We also plan on creating a seamless verification system that can verify the reliability of scientific articles submitted to streaMLined pandemic directory. This provides other scientists with the opportunity to contribute to our project and share resources that can benefit and be used by everyone. With these additional components, our web app will become a more complete directory of pandemic resources.

StreaMLined pandemic directory can be marketed to governments, organizations, and individuals around the world, as the resources and data available is appropriate for all audiences. Our web application is cost-free for anyone to use, and there are no creation or maintenance fees associated with it. If we decide to continue managing our web app in the future, we may need to receive funding to cover the costs affiliated with expanding and operating our project.

References Used:

Technical references(codes, github repo of libraries you used, etc) ML section inspiration:https://github.com/parth57/Car_Price_PyWebIO/blob/master/app.py Dataset: https://github.com/owid/covid-19-data/blob/master/public/data/owid-covid-data.csv PyWebio: https://github.com/pywebio/PyWebIO Scikit learn: https://github.com/scikit-learn/scikit-learn Pandas: https://github.com/pandas-dev/pandas/ Numpy: https://github.com/numpy/numpy Pickle: https://github.com/python/cpython/blob/3.9/Lib/pickle.py Flask: https://github.com/pallets/flask Seaborn: https://github.com/mwaskom/seaborn Matplotlib: https://github.com/matplotlib/matplotlib Statsmodel: https://github.com/statsmodels/statsmodels

Figma Prototype: https://www.figma.com/file/7eRbqb2ruoEA7mPClbKGhH/streaMLined-Pandemic-Directory?node-id=1%3A2 Web App Typeform: https://4tafetccp9s.typeform.com/to/nYwhdeep

Research references(news/journal articles)

Works Cited Bickley, Steve J., et al. How does globalization affect COVID-19 responses? 20 May 2021. 15 August 2021. Dodds, Walter. Disease Now and Potential Future Pandemics. 3 December 2019. 15 August 2021. Gussow, Ayal B., et al. Prediction of the incubation period for COVID-19 and future virus disease outbreaks. 30 November 2020. 15 August 2021. Marín-Hernández, Daniela, Nathaniel Hupert and Douglas F. Nixon. The Immunologists’ Guide to Pandemic Preparedness. February 2021. 15 August 2021. Mason, Diana J. and Christopher R. Friese. Protecting Health Care Workers Against COVID-19—and Being Prepared for Future Pandemics. 19 March 2020. 15 August 2021. Moore, Sarah. The Future of Pandemics. 27 April 2021. 15 August 2021. Pearce, Joshua M. A review of open source ventilators for COVID-19 and future pandemics. 30 April 2020. 15 August 2021. Ready. Pandemics. 24 June 2021. 15 August 2021.

+ 2 more
Share this project:

Updates