Answer by Francis Chen:
First, I'd like to say: data visualization peeps and aspiring entrepreneurs, take note!
Carlo Ratti of the MIT Sensable Cities Lab () has done some amazing work in terms of data visualization and smart cities. Check out their website, because they have been cranking out projects for about several years now.
The MIT Sensable Cities Lab does a lot of work on using "a vast system of cameras, communication devices, microcontrollers and sensors over our environment, enabling entirely new ways to imagine, monitor, and understand our cities."
I wrote a similar post here about some of their work:
Here's a video:
They've recently opened up a research lab, in collaboration with the National University of Singapore: .
However, from past experience, lots of the data-driven, smart cities movement has been very small scale and focused on a few institutions.
From my ala mater, UC Berkeley, UrbanSim is a software simulation tool used to model transportation, land use, economic development, housing, etc. for metropolitan planning organizations in the US.
It definitely doesn't have the real time feel of MIT Sensable Cities (they don't depend on cell phones or cameras for example), but UrbanSim is definitely doing a lot in terms of using DATA to predict and measure how cities will look and be, as well as function in the next twenty or thirty years.
Given that we're in an era of climate change and sustainable cities, these sorts of softwares and technologies will become increasingly important in making sure we can measure how healthy our actions our to the planet as well as ourselves.
I see these as good business/market opportunities as city governments have to plan out how to make their cities livable, and these models will make sure cities are planned accurately based on the right data.
However, there already is a lot of work that has been done in traffic/transportation planning.
The question is how to make this data and software more accessible to the general public.
Most of their software is only available to transportation/traffic planning engineers, and to a certain extent, some urban planners. In addition, much of the user interface is very technical and intimidating, which is a turn off for a lot of people (and thus people are intimidated by lots of urban planning).
Lots of traffic modeling software looks like this (lots of transportation planning is used like this to plan for future infrastructure, by the way)
But, for the average person, it's really hard to relate to this sort of work. For somebody who's interested in the carbon emission output of this, or the impact of travel time in terms of traffic, times spent with family, housing costs, etc.
The Center for Neighborhood and Technology (), an organization based in Chicago, has done an amazing job in collecting US Census Data from all metropolitan areas across the US, measuring housing and transportation costs, relative to income.
This issue is really important, especially in areas like the San Francisco Bay Area, where housing/transport costs are really hard and can limit how much disposable income a family could spend on, lets say leisure, recreation, health, education, food, etc.
Here's a link here:
Columbia University also has a research institute that deals with data driven planning housed under the School of Architecture and Planning, known as the:
Measuring CO2 Levels across Beijing during the Olympics.
However, I didn't see that many actual visuals in their website.
Finally, to give homage to a Bay Area native, lets not forget the indie data visualization experts; people who do data driven urban planning in their free time.
Eric Fischer has an amazing set of work on mapping/data visualizations on his flickr:
Mapping local/tourist activity in New York City using Flickr Geotags
To mapping pedestrian activity from BART stations to people's homes/work:
According to Fischer:
"Routes are approximate: shortest path from the trip origin to the station along TIGER (2010) roadways. Diagonals are especially suspect. BART passenger trip origins from the 2008 Station Profile Study, snapped to the nearest street corner."
P.S. This data is easily available on BART's website:. The routes he made using Geographic Informations Systems Software (likely ArcGIS)
Race and ethnicity maps of Los Angeles:
I also subscribe to this YouTube channel which shows how frequent public transport is throughout major American and European cities. These were made using Google General Transit Feed Specification (GTFS) data. I imagine lots of programming was involved in this:
Here's a day in the life of Washington DC Transit:
In terms of government institutions:
The City of San Francisco MTA recently created a smart parking program called SF Park, which basically prices parking by supply and demand (coined by parking guru and urban planning Professor Shoup from UCLA) based on sensors that have been installed throughout the City of San Francisco's off-street parking and parking structures. This is one of the most largest scale projects of its kind in the US, and was funded by the federal government.
For example, if a particular street corner does not have a lot of parking spaces, the price will increase by $0.25 or $0.50, depending on the amount of spaces left. Meanwhile, a street corner with lots of parking spaces will decrease its prices by the same amount (there's an algorathim which I'm not so sure how it works). Supply and demand, pricing parking at market prices, in action.
Given that parking takes up a lot of urban space in many urban/suburban areas, and urban space and parking spaces are limited, as well as a need for our cities to encourage more usage of public transit, this is a great way of using data to solve problems.
There is a lot of parking occupancy related data available on the SF Park website:
Here's a video explaining the process:
Here's the interface in detail:
Stamen Design (), based in San Francisco ( , please comment!) which does a lot of visualization work (not just cities; their clients have included MTV, MSNBC, SF MOMA, Schwab, Adobe, AirBnB, BMW, and MoveOn):
Mapping Google, Apple, eBay, Facebook, etc. buses from SF to Silicon Valley: (See the process explained here)
I have more examples, but they'll come as the discussion progresses.
Several things to conclude with, from my experience and research:
- These projects require cross-collaborations between different organizations and institutions. Academic institutions such as MIT and UC Berkeley for example have a great knowledge set of people/researchers who have a large knowledge set in these areas. In addition, lots of government / policy think tanks and other similar organizations are dying for good, accurate data (as well as how to measure and use it) to plan their cities to be more livable and healthy. Finally, the private sector is teeming with experts in data visualization/scientists/designers, software engineers, as well as people who are focusing on improving sensor/GPS technology.
- It is not enough to know how to code/program/design or collect/visualize data. I really emphasize this particular point because I believe that a lot of particular social/political/economic issues will be ignored if the entire team of data visualists consist of just those experts. For example, I've seen a lot of data related to transportation because there has been a wealth of research on how to measure transportation (how frequent is the bus running, how accessible is transit from your house, etc.), but not enough data visualization relating to issues relating to: food security, health/exercise, the impact of restaurants/local food businesses in a local economy, the impact of parks in improving health and community friendships, etc. This sort of perspective requires a "very non-programming perspective"; in other words, hire economists, urban planners, architects, anthropologists, policy researchers, small business owners, etc. into your team!
- Not all data is available. Government institutions don't release a lot of data to the public, or some data needs to be collected.
- Lots of data is available, but it is left very unorganized and very un-user friendly. There is actually a lot of data readily available (i.e. think government websites, with those large PDFs).
- With that being said, sensors/cameras/GPS, and other alternative means to collect data will becoming increasingly important in the future.
- Hubs/innovation centers for these organizations will become increasingly important to provide an environment.