Looking Forward to 2021

December 31, 2020 by

Andrew Damico; Brennen Chadburn; Michael Purvis; Cameron Moore and Michael Martley in Partner POV

To close out this year, we asked our partners and associates what they were most looking forward to in data science and analytics in 2021.

I’m interested to see progress in building causal reasoning into machine learning models. Recent work on a multivariate forecasting problem led me to thinking about how predictive modeling could benefit from establishing complex casual relationships. Machine learning models are extremely good at mapping relationships, but I often find minor features that are contextually important get sidelined for more salient features.

I ended up running across an article on “Algorithms for Casual Reasoning in Probability Trees” (Genewein et al.) which provides an intriguing proposal for building causal reasoning into discrete probability trees. Establishing casual reasoning seems an important attribute of a model that could better predict context dependent outcomes. I’m excited to see how this develops in the next year and what impact it might have in forecast modeling.

https://arxiv.org/pdf/2010.12237.pdf

2020 has illuminated many new areas of opportunity for technology with particular emphasis on data science and analytics. Almost overnight it became more imperative than ever before for our heroic front-line emergency first responders and healthcare professionals to have on-demand access to data and trends at any time and in any location. Data as a Service (DaaS) can help address exactly this need. Beyond just storing data in the cloud, DaaS adds efficient and scalable data processing and analytics, mitigating the need for devices to have powerful hardware or a fast internet connection to be useful in the field. I would love to see more rapid and widespread adoption of Data as a Service across all industries.

Michael Purvis

General Partner

Data Engineering & Cloud Infrastructure
Onboarding and Practice Development
Business Operations & Strategy

Andrew D’Amico

General Partner

Business Operations & Strategy
Business Analytics Practice

In 2020 the new “Space Race” began, but unlike the space race of the 1960s private companies such as SpaceX and Virgin Galactic are at the forefront of exploration. NASA’s Artemis Program, the first US lunar landing program since the Apollo 17 mission in 1972, has also picked up speed and funding with touchdown aiming for 2024. These active programs will spurn innovation ultimately reducing the cost barrier to entry for new entrants in orbital GIS satellite research. Northwestern Analytics uses data from several satellites for client projects. I predict the advances in space technology due to the pragmatic needs of these missions will increase our access to quality GIS data and greatly aid our models.

I am most excited for the emerging intersection between data science and the social sciences. This year in particular has shown the importance of mental health, including the proper and affordable access to mental health resources. Data science techniques such as natural language processing and predictive analytics can help to explore societal trends and assess the need for mental health resources and treatment options. This, in turn, can allow social workers and psychologists to better understand and meet the needs of the general public. Data science may also illuminate how the shift towards an increased technological world and a more prominent online presence has impacted mental health, as well as outcomes from various treatments that consumers have tried. My hope is that eventually these lessons learned from the data can be used in a preventative manner to promote better mental health outcomes.

Cameron Moore

Associate

Gis Practice, Computer Vision Practice

Michael Martley

Associate

Gis Practice, Computer Vision Practice

I would like to see more use of prescriptive analytics. Prescriptive analytics can be built off of predictive analytics, but also optimization. There are various approaches to optimization to produce options that the stakeholder can review for decision-making. A simple example of this is supply chain management, where several scenarios can be tested using mixed-integer or binary programming to satisfy scenario requirements, such as: maximizing revenue, maximizing profit, or taking care of customers’ supply needs based on their category (small, medium, large).

About the Author

Andrew Damico

Managing Partner, Operations and Project Management at Northwestern Analytics Partners LLLP | Website

Andrew is a General Partner at Northwestern Analytics and is the managing partner responsible for Business Operations and Project Management. Andrew holds his Project Management Professional certification and also serves as the Chief of Staff of Business Outreach for the Project Management Institute Chicagoland Chapter.

About the Author

Brennen Chadburn

About the Author