Volunteer Position or Internship: Human-Mobility Spatial Modeler

Would you like to be part of an ambitious and innovative team working on cutting-edge projects to help humankind and to make a positive difference in the world?

We are looking for a volunteer or intern to join our team and help us develop our product for COVID-19 Risk Analysis. This is an unpaid position.

Responsibilities:

• Design, implement, and evaluate a parameterization of flow of people between regions for an epidemiological model on COVID-19 risk and spread.

• Identify, compare, and acquire relevant geospatial data sources (if needed)
• Process relevant data into a form convenient for modeling or further analysis (if needed)
• Analyze actual inter-region mobility data and/or related data to formulate the parameterization
• Use historical infection data to constrain and refine the new parameterization
• Validate new parameterization
• Evaluate new parameterization within the greater model using additional historical infection data
• (possibly) extend model to include demographic groups with differing mobility and the degree of interaction between different demographic groups

• Document progress
• Collaborate with the team

Requirements

Experience performing data analysis in Python and/or R

Geostatistics knowledge (strong background in statistics preferred)

GIS/Geospatial database experience

Graph theory (particularly Network theory) knowledge strongly preferred

Familiarity with various Geospatial tools & technologies, such as PostGIS, GDAL, OGRSQL, CGAL, and Python modules: scikit-mobility, movingpandas, GeoPandas.

SQL knowledge is a plus

Team spirit

Good problem-solving skills

Location

Fully Remote

Term

The duration & hours are negotiable.

About Quantum Risk Analytics, Inc.

We are a 501(c)(3) federally-tax-exempt charitable non-profit research organization founded by MIT alumni.

Our mission: To help mitigate the COVID-19 pandemic and future epidemics by developing and deploying robust risk assessment tools that predicts real-world infection, long-COVID, severe disease and mortality risk, thus empowering individuals, groups, and organizations to make better risk-informed decisions about their activities. Our larger mission is to address emergencies with analytics, and we will be branching out into other domains in the future.

Pandemonium is a COVID-19 Risk Assessment & Projection Model being developed by Quantum Risk Analytics, Inc. as its first product.

The basic framework of the model will allow for an arbitrary degree of demographic and spatial refinement where data is available as well as allow various submodels to be incorporated into the overall model making this a very extensive and extensible framework that can be used by researchers and others. (For more about our model design, please refer to the draft proposal on our website.)

Some key technical attributes of our model framework so far:

• Stochastic (draws are made from probability distributions, giving distributions as model results and thus statistical metrics such as confidence intervals)

• Regional & Compartmental (each region’s population is categorized into distinct compartments, the evolution of which is modeled)

• Individualized (allows specified individuals to be modeled in conjunction with the larger epidemiological model)

• Heterogeneous (epidemiological parameters can vary over time and region)

• Spatially-Hierarchical (smaller regions within larger regions can be modeled where data is available for greater refinement and specificity)

• Dynamic (allowing for specific travel history/plans & commuting pattern changes over time)

• Demographically-structured (differences among demographic groups based on age, race, etc., can also be modeled where there is data available)

• Incorporates micro mechanistic submodeling (models the physical mechanics of transmission in micro scenarios, fully integrated with the epidemiological model)

• Incorporates vaccinations (the effectiveness of vaccines as provided by submodels are utilized).

• Modular (allows for various submodels to be developed that are narrow in scope but still benefit from being connected to the larger framework and its other submodels, epidemiological model and inference capabilities)

• Machine Learning built on the Pyro Probabilistic Programming Language (PPL), using Markov Chain Monte Carlo (MCMC) and other statistical inference techniques. (This allows the strongest inference practical from imperfect, missing and limited data, and forecasts into the future.)

We continue to develop this model to include additional features, information & detail. We plan to make our model framework open-source around the same time that we launch our risk assessment app for general use.

Quantum Risk Analytics, Inc. is a 501(c)(3) non-profit organization founded by MIT alumni, bringing together talented, motivated people from different countries and with diverse backgrounds. This has been an all-volunteer effort with volunteers from 3 continents. We want to contribute our skills for the greatest good of the world, devastated by the COVID-19 pandemic. We want to help as many people as possible, including and especially the more marginalized, both in the US and around the world. This is why we wish to make the service available to everyone for free and why Quantum Risk Analytics, Inc. is a non-profit organization. If you have the interest, commitment, and ability to contribute your skills while you gain experience and knowledge working on a real-world product, please apply!

How to Apply

Please, E-mail us with your resume and a brief cover letter.