The Tech - Online EditionMIT's oldest and largest
newspaper & the first
newspaper published
on the web
Boston Weather: 75.0°F | A Few Clouds
Article Tools

Earlier this semester, Provost Martin Schmidt and the Deans of the five schools announced the establishment of the Institute for Data, Systems, and Society, headed by Professor Munther Dahleh. This exciting new entity aims to “address societal challenges using analytical tools from statistics and information and decision systems,” and will officially launch on July 1.

While the inaugural letter and subsequent news coverage of IDSS outlined the entity’s high-level objectives and structure, little has been said about the proposed Center for Statistics. In comparison to peer institutions such as Stanford, Columbia and Berkeley, which have long-established statistics departments, the current structure of statistics at MIT is highly fragmented.

From my own experience, I have enrolled in several well-taught statistics courses in the Civil Engineering, Electrical Engineering and Computer Science, and Brain and Cognitive Science departments. However, given the different approach assumed by each department, it is challenging for students to progressively build a unified sense of sophistication and maturity in statistical thinking.

According to the proposal submitted by Professor Dahleh, The Center for Statistics plans to address this issue by offering coherent, centralized programs in statistics. Degrees at the PhD and Master’s levels are in their planning stages, and a full proposal for an undergraduate minor has already been submitted to the Committees on Curricula and Undergraduate Programs. While establishing world-class graduate level programs is a key priority, it is equally important for the visionaries of IDSS to develop a culture of statistics that interests and engages members of the MIT community from various academic backgrounds.

New introductory courses in Electrical Engineering and Computer Science are already reflecting this goal, such as 6.0002 (Introduction to Computational Thinking and Data Science) and 6.008 (Introduction to Inference). According to Professor Dahleh, a joint effort across all Institute schools is being coordinated to offer new courses in statistics and data sciences. The idea is to extend these opportunities to departments that are not heavily quantitative, such as Political Science, Anthropology, History, and Urban Studies.

The goal is not to create an unconscious statistician out of everyone at MIT. Rather, we should encourage the skill of analyzing data from various fields from a statistical standpoint. Even an elementary probability and statistics requirement (a topic debated at the Institute for over a decade) could lay the essential foundations for incoming freshmen. For example, regression methods yield useful insights when studying data, and they do not require a vast amount of mathematical machinery to get started.

One common misconception is that statistics is a dry subject of mainly theoretical interest. But 21st century statistics is experiencing a revolution, fuelled by the explosion of data and computational power. An inaugural symposium for The Center for Statistics was held at MIT earlier this month, where top academics from across the country discussed their use of statistics in novel areas such as cancer detection and data-driven decision-making in industry. To promote these ideas among the MIT community, Professor Dahleh says the plan is to hold a regular seminar series in the future.

Most of the core elements for a successful statistics center already exist at MIT; over a dozen departments and research centers feature their own flavor of statistics research and course offerings. It will be interesting to see how IDSS integrates various groups across campus in a way that fosters successful collaboration. Creating a hub for sharing ideas between data scientists and social scientists will bring a new dimension to both fields, as well as innovative, real-life research outcomes.

IDSS is a highly ambitious and complex initiative that faces an array of academic, financial, and bureaucratic hurdles. But the project identifies a genuine shortcoming in MIT’s current academic system and outlines a vision for change. Successfully tackling societal issues with rigorous statistical frameworks will cement MIT’s academic and thought leadership, as well as nurture generations of experts in interdisciplinary domains for decades to come.