Between the Poles

Economic value of big geospatial data could reach $700 billion/yr by 2020

McKinsey Global Institute published a report Big data: The next frontier for innovation, competition, and productivity that analyzes the state of digital data, how it is used in different domains in both the private and public sectors, and offers suggestions for policy makers. In addition to looking at big data as a whole, the report examines five areas in detail including personal location data.

Real-time location data generated by mobile phones and other devices has created an entirely new set of location-based services from navigation to pricing property and casualty insurance. Personal location data is a new horizontal application domain that includes industry sectors from utilities to insurance to transportation. The growing smart phone market is creating an exponentially growing volume of location-aware data.  McKinsey sees this domain as a hotbed of innovation that has the potential to create significant value for consumers.  For example, use of real-time  traffic information for navigation will create a quantifiable consumer benefit by saving time spent traveling and fuel consumption. McKeinsey expects the economic value these innovations create to far exceed the revenue generated by service providers.

Economic value

Mobile phones with GNSS chips have created a revolution will continue over the next  decade. This personal location domain offers the potential for huge new value creation over the next ten years that McKinsey estimates at more than $100 billion/yr (about 0.14% of world GDP) in revenue to service providers and as much as $700 billion/yr (about 1% of world GDP) in value to consumer and business end users. Realizing this value will require investment in technology, infrastructure, and personnel as well as appropriate government action.

McKinsey's detailed analysis of the major applications of personal location data finds that, in ten years’ time, these applications could create value of $100 billion/yr or more for service providers  primarily from sales of navigation hardware and revenue from LBS, mobile LBS premiums, and geo-targeted advertising.

McKinsey estimates that by 2020 the benefits (difference between willingness to pay and the cost of applications) that customers, consumers and businesses, will enjoy from location-based applications will reach as as $700 billion/yr. Most of this, about 70 %, will be the consumer benefit from time and fuel saved by using GNSS navigation systems including those with real-time traffic information and mobile LBS applications. The remaining 30 % of the benefits will be through location-enabled applications such as geo-targeted mobile advertising.  Smart navigation applications are estimated to provide $500 billion in value to global consumers in time and fuel saved by 2020. Geo-targeted advertising could represent more than 5 percent of total global advertising spending by 2020.

For comparison, a recent study prepared for Google by Oxera Consulting Ltd in the UK estimated global revenues from geospatial products and services to be $150-$270 billion per year.  The geospatial industry is estimated to be growing by 13% per year through 2016.  These numbers were estimated using a different methodology to McKinsey's. The lower estimate of $150 billion/year was estimated using a bottom-up approach based on the Bloomberg BICS classification of companies worldwide and assumptions about allocating the revenue from companies that provide geospatial in addition to other services such as Trimble.  The $270 billion number was estimated by scaling up the US revenue estimated by Boston Consulting Group to come up with a world estimate.

Personal location data

McKinsey includes sources of personal location data that locate an individual to at least within a couple of city blocks in an urban settingin its analysis.

• Credit and debit card payments from cards processed at point-of-sale (POS) terminals and  automated teller machines. In 2008, there were 90 billion to 100 billion transactions of this type worldwide.
• Mobile phones - the location of such devices can be triangulated from cell-tower signals.  In 2011 there were about 6 billion of these.
• Smartphones with GPS - about 600 million devices in 2010 and the number is growing at about 20 % per year. Many smartphones also have Wi-Fi networking capability, which can be linked to a physical location of a Wi-Fi network.

In the future McKinsey expects indoor location to become more important.  They mention Shopkick, which allows merchants to track their customers from when they walk into a store, and Path Intelligence, which can track foot traffic within malls or amusement parks.

McKInsey's research estimates that the global pool of personal location data was at least 1 petabyte in 2009 and that this pool is growing by about 20 % a year. 

Location data generated by smartphones is growing at double the average rate of growth for all location data.  Asia is the leading region for the generation of personal location data because so many mobile phones are in use there. In 2010 the regions with the largest number of cell phone users were

• China 800 million
• India 650 million
• North America 300 million.

Applications of personal location data

McKInsey has identified three major categories of applications of personal location data.

• Location-based applications and services for individuals including smart routing, automotive telematics, and mobile-phone based location services
• Organizational use of individual personal location data that includes geo-targeted advertising, electronic toll collection, insurance pricing, and emergency response.
• Macro-level use of aggregate location data that includes urban planning and retail business intelligence.

Barriers to use of personal location data

Concerns about privacy and security was identified by McKinsey as a major barrier to future expansion of the use of personal location data. Laws are unclear about who owns that data and how it can be used.  McKinsey believes that a policy and legal framework that clearly describes the permissible and prohibited use of these data would be beneficial for all stakeholders.

For policy makers, McKinsey sees that a priority is incentives for innovation are in place including developing an up-to-date framework for intellectual property rules and rights, funding R&D in potential breakthrough areas, and ensuring that the infrastructure, including spectrum policies, is optimal.

Methodology

By looking at the value chain of location-based applications, McKinsey examined both the amount of data generation and the potential to create value.  McKInsey estimated the amount of personal location data generated in each category of application orsource device and in different geographic regions using the installed base of devices, usage behavior, and the frequency with which data are generated. To assess the potential value that data can generate, McKinsey estimated the economic value of each application to end users—individual consumers, enterprises, and government organizations.

Location-enabled Society: Mapping tweets using real-time big data technology and privacy

In a very appropriately titled session Taking Society's Pulse in Real-time at the Creating the Policy and Legal Framework for a Location-enabled Society conference in Boston, one item really got people's attention focussed on the issues that arise from combining location with personal data. TweetMap is a technology that visualizes a large number of tweets on a map of the world. This is not the first time I have seen tweets mapped.  But what really got people's attention is that a speaker showed how to drill down to your neighbourhood and see what the people next door are tweeting about.  Of course you can do this without a map by simply following your neighbors on Twitter.  But somehow mapping it makes it seem more intrusive, even invasive to some people. In another demonstration, Tweetmap was used to show how a person sailing in the Caribbean could be tracked as they tweeted from their boat.

To be tracked geospatially on Twitter requires that you explicitly agree to geolocation tracking for Twitter on your iPhone or Android device.  For example, on an iPhone the geolocation option allows you to make your location known to others. Apps such as Facebook or Twitter for iPhone can report your location as well as the messages you post.  The iPhone requires that these applications obtain your permission before they can be enabled.  Apparently only about 2-4% of smart phone users, mostly younger folks, agree to this.  After seeing a Tweetmap drill down, I would expect that some people might have second thoughts about agreeing to the gelocation option. And in the not too distant future this will be possible in real-time on inexpensive hardware with open souce software.

TweetMap is an instance of MapD, a massively parallel database platform being developed through a collaboration between Todd Mostak, (a researcher at MIT), and the Harvard Center for Geographic Analysis (CGA).  Currently the tweet database starts on 12/10/2012 and ends on 12/31/2012 and contains 125 million tweets.  The database can be queried by time, space, and keyword.  The ultimate objective is to be able to query billions of tweets live in real-time.  In this case real-time means from tweet to tweet on a map in under a second.

MapD is a general purpose SQL database that can be used to provide real-time visualization and analysis of very large data sets.  MapD uses commodity Graphic Processing Units (GPUs) to parallelize compute intensive tasks such as querying and rendering very large data sets on-the-fly.  MapD runs on inexpensive hardware ( ~$1000) and is intended to be open source.

Geodesign Summit: Carl Steinitz on the early days of geodesign and organizing education for geodesign

Last night Carl Steinitz, Professor of Landscape Architecture and Planning at the Graduate School of Design at Harvard, gave two presentations.  The first addressed the queston of how to organize education in geosdesign and was the result of Carl being invited to advise a specific university (which remained nameless).  Last night's talk was the actual presentation he gave at the university at the culmination of a week's visit to the university.  The second was a personal history of the early days in the development of geodesign, primarily in the 60s.

Organizing education in geodesign

Geodesign involves bringing together the design professions (architecture and engineering), technology, the geographic sciences, and wha Carl calls the people of the place, the folks that are going to live in the result of a "geodesigned" environment.   A major challenge arises from scale.  Scientists approaach things from the universal anf global, designers from the local, even at the level of  one building or one parcel. 

Carl made the case that the most important thing that all students of geodesign, regardless of their background, need to learn is how to collaborate in a multi-disciplinary project environment involving architects, engineers, IT folks, and folks from geographically-oriented sciences.  He made specific recommendations on how to do this, not generally applicable to all unversities but tailored to the specific unnamed univesity he had been asked to advise.

Early days of geodesign

There is a fascinating book by Nick Chrisman called How computer mapping at Harvard became GIS that describes the early days in the development of geospatial technology and science. 

Carl Steinitz was part of these early developments.  Last night he gave a personal perspective on the early days of geodesign. 

From a technical perspective the developments I found most fascinating were the first computer geographic computer graphics that was created by hooking an IBM Selectric typewriter to a computer and using overprinting to create an approximation of grey scale thematic maps. The program to do this, SYMAP,  was developed by Howard Fisher in 1963.  When the Harvard Lab for Computer Graphics was founded in 1966, Howard Fisher was its first director.  Carl also described some of the ways colour printing was simulated in these early days, and early 3D imagery.

An interesting recollection in this fascinating talk was that computing was introduced into the design course via tutorials as and when needed, not as a separate course.

Carl described the development of geodesign methodogies such as moving from traditional programming languages in the direction of what became map algebra, spatial analysis, automating aspects of design and using linear programming to optimize design.  As Carl said they did it all, though some of it was too expensive at that time to be of practical use.

In memory of Wolfgang Bitterlich, one of the early innovators in geospatial analytics

Last week a memorial event was held in Ottawa for Wolfgang Bitterlich, one of the early technical innovators in spatial analyics.  Organized by his friend Mike Comeau, the memorial brought together a large group of Wolfgang's friends and colleagues including many that I have known from the early days of DPA and Tydac;  Laura Bobak, the daughter of Wolfgang's long time partner Esther Bobak and whom together with her sister Nadya Wolf regarded as his own daughters, Giulio Maffini and Maryjane Maffini, Richard Higgins, Louis Burry and his wife Sandra, Anne-Marie Hogue,Terry and Iris Moloney, Adla Worobec and Graham Stickler, Wendy and Dave Branson, Peter and Erin Kuciak, Kate Dickinson, Bruce Thomas, Rob and Beth VanWyngaarden, Bob Madill, and Chris McBean, Fred and Joanne Meth, and Klaus and Judy.  Quite a few years ago Wolf moved to Los Angeles, so this was the first opportunity for many of us in Ottawa to meet Wolf's wife Nidia, and her family, Elizabeth, Danny and Priscilla from Los Angeles.

Rick Higgins who was master of ceremonies, provided an opportunity for other people also to speak about their recollections of Wolf - Fred Meth, who had known Wolf since 1971; Giulio Maffini, the other half of the dynamic concept/technical team that conceived and implemented Tydac's SPANS (Spatial Analysis System); Laura Bobak, who gave a touching perspective on what Wolf meant to her as a surrogate father beginning at age eight; Maryjane Maffini who read a letter from Michael Simmons who had been with Wolf at Tydac; myself; Danny, Nidia's son; Rick Higgins who told a very funny story about Wolf as the Russian ambassador in Sri Lanka; Adla Worobec; Louis Burry, Wolf's right hand man who worked with Wolf on the technical side for many years in the 80's and early 90's; Doron Nussbaum who also worked with Wolf at Tydac; Judy and Klaus, who played chess with Wolf in Vancouver; and finally Maryjane read a letter from Fiona Smith in Australia.

SPANS

To put Wolf''s contribution in context, Canada has been an early leader in geospatial technology.  In the mid 1960’s, the Canada Geographic Information System (CGIS), written in IBM 360 assembler and running on OS360 was developed under the aegis of the Department of Forestry and Rural Development.  By the late 1960's and 70s, and through the 1980's the Canada Land Inventory (CLI) was managing a geospatial database containing 2.5 million square kilometers of land and water. (The CLI database is still available on GeoGratis).

Richard Higgins founded DPA Consulting in October, 1973 in Halifax.  Giulio Maffini joined DPA in 1975 and together they managed to attract Wolfgang from the Technical University of Nova Scotia to join DPA, aslo in 1975.  As a result of the combination of Giulio's conceptual and Wolf's technical genius, many DPA projects were based on innovative modeling and simulation technologies.  In the early 1980's a few of these projects involved land use planning, for example, for Dartmouth, Nova Scotia and Sri Lanka.  Land use programming requires geospatial analysis, and Wolf implemented an early, but innovative, spatial analytics application based on a grid to support these projects.  In those days the only alternatives for printing graphics were very high end printers like Applicon and Calcomp, pen plotters, and dot matrix printers such as Printronix that allowed you address individual pixels.  I joined DPA in 1981 and to my recollection all of the graphics included in the Dartmouth and Sri Lanka reports were produced on a Printronix printer.  At this time, to give some context for the state of geospatial technology, IBM's GFIS system was storing spatial data in a hierarchical database on OS360 and ArcInfo, written in Fortran, was running on mini-computers.

The early 1980's were an inflection point in computing.  I had bought and assembled my first microcomputer, a Heathkit H89 with a Z80 CPU, in the late seventies and I had just successfully completed a DPA survey project using my H89 at considerably less than using an IBM360 timeshare would have cost.  It was obvious to me and to Richard and Giulio that the microcomputer dramatically reduced the cost of entry compared to mainframes and mini-computers making it possible for even small companies to develop and market products for the new platform.  

In 1984/1985 Richard and Giulio, projecting that the spatial analytics market had a large potential and seeing the rapid expansion of the microcomputer user community after the incredibly successful introduction in 1981 of the IBM PC, founded Tydac Technologies to develop and market a spatial software product that became known as SPANS. 

The initial Tydac team was the combination of Richard's business acumen and Giulio's conceptual and Wolfgang's mathematical genius.  Later Louis Burry brought his programming and system architecture expertise to Tydac to complement Wolf's mathematical genius. 

The most innovative element of the SPANS system was the use of quadtrees, a mathematical approach used by Wolf to provide an efficient grid-based foundation for spatial analytics. 

Another major innovation was potential mapping, known as POT MAP to SPANS users, and as multi-factor suitability analysis to the broader geospatial community, that allows you to model and display spatially land use suitability for agricultural, foresty, industrial, and commercial projects.  The SPANS modeling language allowed users to perform very sophisticated suitability analyses involving many variables.  SPANS users believe that even today no GIS has reached the level of spatial analytics that was possible with the SPANS modeling language.  SPANS originally was developed and ran on DOS with high end graphics cards from companies such as Number Nine.  Later it was migrated to OS/2.  The first SPANS product on Windows was SPANS MAP.

I joined Tydac in 1987 and by the early 1990's Tydac was one was one of the top 10 GIS companies in the world with more than 3000 systems installed in over 60 countries.  In the December 1992 issue of InfoWorld, there was a very interesting review of several of the geospatial products that were available at that time including Atlas GIS 2.0, MapInfo for Windows 2.0, and SPANS GIS 5.0.  The review was very perceptive,

"Tydac Technologies Corp's Spatial Analysis System (SPANS) GIS is the most powerful program in this comparison.  In addition to its capability to incorporate both raster and vector images into maps, this OS/2 program offers an extremely strong, if difficult-to-learn, set of analytic tools. ... Surprisingly, SPANS GIS lacks some of the basic features one would expect in such as high-end program, such as address geocoding and a traditional database interface.  And the program almost completely lacks formatting features required to print high-quality map presentations.  For these tasks, SPANS GIS relies upon a lower end sibling, SPANS MAP, which is capable of loading and altering maps from SPANS GIS."

As I was the project lead for SPANS MAP, I knew exactly what the technical issues were, but what is perhaps only clear in hindsight is that SPANS was an innovative spatial analytics engine, way ahead of its time, capable of the most sophisticated spatial analysis.  To categorize it as a general purpose GIS was misrepresenting it and its capabilities.

To continue the saga of Tydac, in 1991 Flavio Hendry founded Tydac AG in Bern, Switzerland, so the Tydac name continues to this day.  In fact a few years ago, Flavio asked me to come to the Tydac AG User Group meeting in Küsnacht to speak on the topic "20 Years of GIS: From SPANS to Google".  The SPANS code base was acquired by PCI, the company to which Louis Burry moved, and is incorporated in PCI's Geomatica.

Alll of us who were involved with SPANS recognize that Wolf's innovative technical genius provided the foundation for an, even to this day, amazingly innovative geospatial analytical engine that moved goepatial analytics far ahead of the simple mapping applications that were typical of GIS products in the 1980's.  But many of us remember Wolf  not just as  colleague but as a friend.  Laura's recollections of Wolf were so poignant at the memorial event.  Giulio's and others' recollections expressed such a sadness about Wolf's leaving us because he was a friend.   My wife Ellen and I often went swimming with Wolf at Blanchard Beach on Meech Lake, his favourite swimming location.  We also remember having a great time touring the lock sites on the Rideau Canal with Wolf and Esther.  But what remains in my memory, and it was expressed so well in Laura's recollections, was Wolf's incredible loyalty to those near and dear to him.  He was a great innovator and for many of us a good friend.