UVA to Set Up School of Data Science With $120 Million Gift

[JacksinPA]

https://www.wsj.com/articles/uva-to-set-up-school-of-data-science-with-120-million-gift-11547823600

Largest donation in school history being used to establish program in fast-growing industry

The University of Virginia will establish a school of data science with a $120 million gift—the largest in its 200-year history, the school said Friday.

The funds come from Jaffray Woodriff, a 1991 graduate of UVA’s McIntire School of Commerce, through his Quantitative Foundation. His wife, Merrill Woodriff, who received both an undergraduate and master’s degree from UVA, is the foundation’s director.
=======================================
Sounds like they are putting this money in the right area.

As a side not about data science, the boyfriend of a former neighbor's daughter just accepted an internship at Microsoft in the Seattle area. He just graduated from RPI, a top engineering & data science college, & writes computer code. Their offer: $80K/year, free housing & a car. That's just to start.

 

[Xelor]

https://www.wsj.com/articles/uva-to-set-up-school-of-data-science-with-120-million-gift-11547823600

Largest donation in school history being used to establish program in fast-growing industry

The University of Virginia will establish a school of data science with a $120 million gift—the largest in its 200-year history, the school said Friday.

The funds come from Jaffray Woodriff, a 1991 graduate of UVA’s McIntire School of Commerce, through his Quantitative Foundation. His wife, Merrill Woodriff, who received both an undergraduate and master’s degree from UVA, is the foundation’s director.
=======================================
Sounds like they are putting this money in the right area.

As a side not about data science, the boyfriend of a former neighbor's daughter just accepted an internship at Microsoft in the Seattle area. He just graduated from RPI, a top engineering & data science college, & writes computer code. Their offer: $80K/year, free housing & a car. That's just to start.

Red:
Interesting, if a bit odd.

Not clear why an school, as opposed to a program, of data science is a needed thing. These days, pretty much everything is driven by data analysis -- public policy (unless one is a Trumpkin), psychology, sociology, economics, operations management, and scores more. The thing about data is that using it aptly to design and develop solutions depends not only on understanding a process' data, but also on having a keen comprehension of the process itself, the latter being essential to knowing at what level to obtain, partition, apply, heed and disregard the learnings the data can provide. In other words, as with everything, context is everything, and that is so with data and its uses.

Because of the context-critical nature of data use/analysis in all disciplines, the data and process analysis compartmentalizations approach most aptly germane to evalue, say, factory floor operations is not the same one that works for, say, finance or IT operations, or even the sales process. Adroitness with a given process is something that comes not from being a data expert, but from being a process expert, a subject matter expert who has quantitative analysis acumen. (That's not to say a data expert cannot obtain the process comprehension; however, it's a hell of a lot easier to master the process and then make sense of the data it produces than it is to grasp the data and "back walk" it to arrive at process comprehension.)

That said, I'm not poo-pooing the notion of a school of data science. I just find it a bit bizarre given the multidisciplinary nature of data and the multifariousness of aptly using it. Even so, I'm sure the deans will figure out how to overcome the dilemmas presented by those "disconnects." To be sure, if one approached a university and told them that in exchange for establishing a school of effluvium arts and sciences, one'd give them $120M, the school would find a way to abide one's wishes.

 

[JacksinPA]

Red:
Interesting, if a bit odd.

Not clear why an school, as opposed to a program, of data science is a needed thing. These days, pretty much everything is driven by data analysis -- public policy (unless one is a Trumpkin), psychology, sociology, economics, operations management, and scores more. The thing about data is that using it aptly to design and develop solutions depends not only on understanding a process' data, but also on having a keen comprehension of the process itself, the latter being essential to knowing at what level to obtain, partition, apply, heed and disregard the learnings the data can provide. In other words, as with everything, context is everything, and that is so with data and its uses.

Because of the context-critical nature of data use/analysis in all disciplines, the data and process analysis compartmentalizations approach most aptly germane to evalue, say, factory floor operations is not the same one that works for, say, finance or IT operations, or even the sales process. Adroitness with a given process is something that comes not from being a data expert, but from being a process expert, a subject matter expert who has quantitative analysis acumen. (That's not to say a data expert cannot obtain the process comprehension; however, it's a hell of a lot easier to master the process and then make sense of the data it produces than it is to grasp the data and "back walk" it to arrive at process comprehension.)

That said, I'm not poo-pooing the notion of a school of data science. I just find it a bit bizarre given the multidisciplinary nature of data and the multifariousness of aptly using it. Even so, I'm sure the deans will figure out how to overcome the dilemmas presented by those "disconnects." To be sure, if one approached a university and told them that in exchange for establishing a school of effluvium arts and sciences, one'd give them $120M, the school would find a way to abide one's wishes.

I'd say, as a guess, that the multifaceted nature of IT - from circuit boards to complex algorithms - requires a blanket descriptor like 'data science.'

 

[Moon]

https://www.wsj.com/articles/uva-to-set-up-school-of-data-science-with-120-million-gift-11547823600

Largest donation in school history being used to establish program in fast-growing industry

The University of Virginia will establish a school of data science with a $120 million gift—the largest in its 200-year history, the school said Friday.

The funds come from Jaffray Woodriff, a 1991 graduate of UVA’s McIntire School of Commerce, through his Quantitative Foundation. His wife, Merrill Woodriff, who received both an undergraduate and master’s degree from UVA, is the foundation’s director.
=======================================
Sounds like they are putting this money in the right area.

As a side not about data science, the boyfriend of a former neighbor's daughter just accepted an internship at Microsoft in the Seattle area. He just graduated from RPI, a top engineering & data science college, & writes computer code. Their offer: $80K/year, free housing & a car. That's just to start.

The company I work for has hired several Data Scientists over the last few years. Challenge we have is making sense of all of the data both ours and our customers systems gather up.

 

[Xelor]

I'd say, as a guess, that the multifaceted nature of IT - from circuit boards to complex algorithms - requires a blanket descriptor like 'data science.'

Perhaps, but I don't see the point of that given that it's already addressed in extant programs.

Unless I misconstrue that of which you write, circuits and algorithms, is already the stuff of an electrical engineering degree (see: course descriptions here) with a couple computer science, statistics and quantitative methods as upper level electives.

 

[JacksinPA]

Perhaps, but I don't see the point of that given that it's already addressed in extant programs.

Unless I misconstrue that of which you write, circuits and algorithms, is already the stuff of an electrical engineering degree (see: course descriptions here) with a couple computer science, statistics and quantitative methods as upper level electives.

I learned a programming language & programming some years ago with an HP RPN programmable calculator. I was into ham radio at the time & I wanted to write a program for tracking ham radio communications satellites. There are 5 trigonometric equations involved. I wound up buying an HP-97 with a printer & magnetic card memory. It took me months to write & debug the program. Used all the 200+ memory registers & had to reuse several memory slots several times to store intermediate results from the different equations.

The basic operation of the program was to initially enter your location's latitude & longitude. A ham radio magazine published the time & longitude for each day's first northbound crossing of the satellite over the equator, so that data was entered next. Then the program was started & ran in a continuous iterative loop, constantly calculating the satellite's position with respect to your station. When the calculator showed that it had risen above your 'radio horizon,' the calculator would print the time, the azimuth & the elevation for each 5-minute span of time so you could point your directional antenna at it.

The final proof that the program was working correctly was entering the data for day 1 & then letting it run overnight until it had printed out the results for day 2. Bingo, it worked as intended.

I don't believe electrical engineering courses get deeply into designing algorithms & writing code to create functioning programs but I have heard that one of the entrance requirements for MIT is writing a program from scratch. Some may.

 

[Xelor]

I learned a programming language & programming some years ago with an HP RPN programmable calculator. I was into ham radio at the time & I wanted to write a program for tracking ham radio communications satellites. There are 5 trigonometric equations involved. I wound up buying an HP-97 with a printer & magnetic card memory. It took me months to write & debug the program. Used all the 200+ memory registers & had to reuse several memory slots several times to store intermediate results from the different equations.

The basic operation of the program was to initially enter your location's latitude & longitude. A ham radio magazine published the time & longitude for each day's first northbound crossing of the satellite over the equator, so that data was entered next. Then the program was started & ran in a continuous iterative loop, constantly calculating the satellite's position with respect to your station. When the calculator showed that it had risen above your 'radio horizon,' the calculator would print the time, the azimuth & the elevation for each 5-minute span of time so you could point your directional antenna at it.

The final proof that the program was working correctly was entering the data for day 1 & then letting it run overnight until it had printed out the results for day 2. Bingo, it worked as intended.

I don't believe electrical engineering [EE] courses get deeply into designing algorithms & writing code to create functioning programs but I have heard that one of the entrance requirements for MIT is writing a program from scratch. Some may.

Red:
I agree that the EE coursework doesn't get into the "esoterica" of code writing. Apologies if my remarks led you to think that's what I had in mind.

It's the electives + the EE-specific courses that make an EE degree provide instruction of the sort you mentioned. Does that mean every EE major will take such programming courses? Of course, not. What it means and what I was getting at by my remarks to which your above post is a reply, is that there is path in the EE major for obtaining undergraduate-level mastery of "the multifaceted nature of IT - from circuit boards to complex algorithms."

Also, though I mentioned the EE major, that's not the only pathway to achieving the outcome you posited. Some of the other ways include:

• Computer science major + EE minor
• Any major + the necessary computer science and EE (electronics) courses taken as electives
  • Courses can be taken at one's four-year institution or there and at a community college

That notwithstanding, as I noted initially, the key thing I don't fully understand is the impetus for a full-on school of data science. FWIW, I spoke with "someone who would know" in the UVA provost's office regarding (1) the factors that distinguish a "school of data science" from the myriad data science curricula in extant programs, and (2) what be the vision for overcoming the experiential and subject matter lacunae that result from teaching/studying data science as an abstraction. The answer I got is that they haven't haven't yet figured out those things; they're working on it.