Rail, Aquariums, and Data

Author: Josh Patterson

Date: May 25th, 2018

This is the 2nd article in a 5-part series supporting the Smart Communications and Analysis Lab (SCAL) and the 2018 Chattanooga Deep Learning Conference.

Economic Impact of Research and Development

We're seeing hub-economy tech companies appear in certain large cities, but not others. These class of technology companies capture a disproportionate and expanding share of value of the industries they disrupt. What is the separating factor in terms of which cities tend to create key technology companies and which ones do not? An interesting correlation is looking at which cities have universities generating large amounts of research as we see in the graph to the right. The basis for our thesis on investing in graduate programs comes from the correlation of economic growth in cities which can be tied to top end universities and publication counts. The quote below from the article goes on to state:

“Both analyses highlight cities in which scientific output is growing. In particular, Beijing, which churned out 0.76% of the global output in 1996, produced 2.74% in 2008 (319,000 research papers). Other fast-growing areas are Tehran, Istanbul, Seoul, Singapore City and São Paulo.”

C. W. Matthiessen, A. W. Schwarz and S. Find Urban Stud. 47, 1879–1897; 2010

In the United States both the San Franciso area ("Silicon Valley") and Boston are well-known hubs of technology companies. We can see in chart above they both rate in hte top 10 in the world in producing research publications.

The article goes on to further examine how citation count of publications can further color the discussion to see how quantity is not always better than quality. In this light we see how cities like Boston and Austin produce high quality research while not always producing the same numbers of publications, but keeping citation count high.
The article goes on to make the observation:

"Take three or four of the best universities in the world, put them in a city with a seaport, and voilà!" says Lobo. But copying the region's formula is quite another matter. How can one city start to emulate another that attracts the most research funding in the United States and has been built up over centuries?”

From the article:
Cities: Building the best cities for science

The quote teases the question about how another city can copy some of the same ideas from successful cities like Boston, or Austin, which tend to "punch above their weight". We'll take a look at the recommendations later on in this article.

Beyond citations, we also see large investment dollars by large technology companies in new research labs clustered in these same cities and emerging regions like Toronto[footnote:Toronto], Canada:

These lab investment tend to cluster in cities that house high-end research institutes that produce specific types of STEM research. In the examples above, we see the investment largely cluster around graduate programs producing deep learning research, as the field of deep learning is largely credited for advances in computer vision, robotics, and advanced sensor analysis (among others).

Correlating High-Output Research Regions with Key Technology Company Formation

Now let's correlate these high-producing research cities with key large technology companies, and trace those back to the ones that were founded from students at the schools.

  • San Francisco Area (Silicon Valley)
  • Boston, MA
  • Seattle, WA
  • Austin, TX
In San Francisco we have more core companies than we can exhaustively list, but setting schools aside for a moment, easy names to throw out that are based in the Silicon Valley area: Along with their class of 2017 IPOs:

So let's take a closer look at how many of these are tied to schools in the region. The key schools[footnote:Schools] in the area are Stanford and University of CA at Berkeley. This region was home to an early wave of key technology companies based out of Stanford:

More recent notable companies that trace their origin stories back to those schools include: Then we have prolific researchers whose careers spanned multiple schools (Berkeley, MIT) and created multiple companies at both school, most notably Dr. Michael Stonebraker:
  • Daniel Abadi, co-founder of Hadapt (acquired by Teradata)
  • Robert Epstein (founder and former VP of engineering of Sybase)
  • Diane Greene (co-founder and former CEO of VMWare)
  • Paula Hawthorn (founder of Britton-Lee, formerly VP of engineering of Informix)
  • Gerald Held (former VP of engineering of Oracle)
  • Joseph M. Hellerstein (faculty at UC Berkeley, co-founder of Trifacta)
  • Mike Olson (former CEO of Sleepycat Software and founding CEO of Cloudera)
  • Margo Seltzer (Professor of Computer Science at Harvard, founder and former CTO of Sleepycat Software)
  • Dale Skeen, founder of Tibco, founder and CEO of Vitria
  • Anupam Bhide, founder and CEO of Calsoft
Combined, Cal Berkeley and Stanford have likely incubated more tech success stories than any other region in the world in terms of pure market cap valuation. When you have 50 years of compounding effects of private and public joint investment in a region (and favorable market trends), we see a region create economic output that is likely unique in the world today (translation: its hard to replicate this easily).

We can go on to examine some of the success stories in the Boston region[footnote:Boston] :

Powerhouse research programs such as Harvard, MIT, and more call the Boston area home. We not only see many companies formed in the area around this cluster of research output, but we also see a lot of companies like Amazon and Microsoft putting labs in this regions as well.

With Seattle, WA, obviously the two most famous technology companies would be Microsoft and Amazon. Jeff Bezos is on record as saying engineering talent availability was a major factor in starting Amazon in Seattle[footnote:Seattle] with the University of Washington as a fertile computer science recruiting ground.

Austin, TX is another interesting case where we don't see as many of the companies formed there, but we see nearly every major tech company with a large office in Austin. Companies founded in Austin, TX Area:

Major companies with large offices in Austin, TX:

While not every single company was founded by 2 people out of a garage in Palo Alto, a lot of these companies have at least part of their origin story connected to research produced by high-performing graduate programs. These people ended up in the area specifically for those graduate programs, and ended up staying close to found key technology companies.

A counter-argument to the relationship between these cities and high output graduate programs might be the argument that these cities have well-funded computer science programs due to the abundance of wealthy tech companies pouring in money; its more likely that these programs feed the region with candidates, pulling in companies, who pour money back into programs in a virtuous cycle. What we do not see are regions that have robust tech sectors without high-producing regional phd programs in computer science.

The compounding and derivative effects of regional graduate research, research labs, and the formation of key technology companies created tend to show yeild over time in certain key cities. We believe investment in specifically graduate research programs is how these ecosystems of technology are booted up for the long-run. Those things being said, these are long-term-class investments, which are not easy to make. Let's take a look at what factors have contributed to success in fostering research output in regions (and then what did not work).

Next: Factors for Fostering Regional Scientific Research Output

Footnotes

  1. Schools: obviously I'm going to get email about the other 20 schools I'm not listing here, but this article got long and I had to focus on a core set. You'll probably email me anyway.
  2. Boston: more than 30 colleges and universities call Boston home
  3. Toronto: Toronto has long had a solid computer science program, but it wasn't until the past 10 years when deep learning exploded did they attract lab investment from key tech companies. Hinton (and others) worked on neural networks for 30 years before getting any attention.
  4. Seattle: Not going to list all of the companies in the Seattle area, although there are many. Microsoft and Amazon tend to flood the area with talent + money and overshadow the University of Washington, but the University was an original attractor for Amazon. Gates had originally located MS in Albuquerque, New Mexico to be close to the makers of the Altair PC. However, he had issues hiring engineers in that area and was from Seattle, so he relocated back to Seattle. The familiarity with the area the ability to hire engineers were amonge the factors for the move.
  5. Google: Google's co-founders met while in grad school at Stanford (developing PageRank together), and eventually moved down the road to Mountain View to build the company.