What are the main areas under active research and development for Geographic Information Science (GISc) i.e. what areas need further R&D?
Some "hot-topics" for GISc may be modelling, simulation, temporal representation.
-
You should be discussing this with your thesis advisor. If the area is not one that he or she is familiar with, you may have to find a co-advisor, or accept that you may not get as much help as you'd like. See multiple Q&A on this at Academia.SE.– mkennedyCommented Dec 15, 2014 at 17:29
Comments disabled on deleted / locked posts / reviews
|
9 Answers
I consider these open and ongoing topics in GIScience:
- implications of user generated content (aka Volunteered Geographic Information Systems)
- geographic effects on social networks
- geographic network analysis
- geographically-enabled agent based modeling
- spatio-temporal structures and analysis
- rapid, interactive experimentation (aka geodesign)
- spatial information infrastructure
- object-based data models for continuous data
- real time and iterative geographic analysis
- analyses on the spheroid
- dataset conflation
- interaction between semantic and geographic search
- mobile mapping and location based services
- human perception of evolving geographic patterns
- implications and algorithms of mixed and augmented reality
-
These look great--but could you provide some references or other support for how you know these are under active research?– whuberCommented Jun 24, 2013 at 17:16
- mobile augmented reality
- geographic data mining
- volunteered geographic information environmental monitoring
- realtime sensor networks
Automatic, yet appropriate, generalisation.
Being able to take high order geometry with a lot of detail and simplifying it for a coarser detail map, without dropping important features, is darned difficult. For example a chain of small lakes visible at 1:50,000 should not be shown at all at 1:500,000, yet the watercourse that connects them should remain visible, and continuous.
Automatic geocoding.
So far as I know, MetaCarta is the only company talking about or providing a service which attempts to automatically georeference any document based on it's content. For example it knows Mark Twain's Tom Sawyer lives along the Mississippi River. This is a rich field and there is a lot of room for more players and implementations.
-
1Sadly last time I checked it also thought the Mississippi ends in France (Orleans). Commented Jun 24, 2013 at 14:57
Big spatial data analytics using open source software for distributed computing such as Hadoop.
There is huge potential for processing massive datasets like high density Lidar data in a distributed computing environment. Berkeley Open Infrastructure for Network Computing (BOINC) is currently an open-source platform for distributed computing. ESRI has already entered the arena by creating Big Data Spatial Analytics for the Hadoop Framework.
-
1+1 You're practically the first, in nearly three years, to offer some references and support for your opinions in this thread!– whuberCommented Jun 24, 2013 at 17:15
Implicit or suggested topology.
wouldn't it be wonderful if the computer noticed that the geometries of layers X,Y & Z were very similar to each other, nearly always following the same trends, and offered to conflate/merge them, or keep the others in lockstep when one is changed?
-
@user19400: the contribution you made should be a comment to the answer rather than an inline edit. Use the [edit xx time ago] link to retrieve your text. While it does continue and support the theme, it is a new thought. It would also be good to point to the results of that research, or the names of the papers, if that is at all possible. Commented Jun 24, 2013 at 16:56
Use of robotics for spatial data collection doesn't seem to be hot - but I think it should be.
Oceans cover most of the earth. Mapping them will require robots.
There's a $7 million prize being offered by XPrize.org.
Human perception and cognition is limited and those limits are becoming increasingly problematic as the volume and variety of information continues to explode in amount and complexity. How can the tools of space and location and representation be leveraged to transform this cacophony of data into pieces understandable, and actionable, to the human mind?
Parallel GIS processing was hot 12 years ago, but seems to have slowly faded. (The link to the "GIS Parallel Architectures Lab" on this page is broken, I wonder if the lab still exists). With so much interest in multicore and cloud, it seems like there should be growing interest in parallel geoprocessing too.
A lot of people say the best way to go parallel is via Functional Programming. That might be a good area, but it seems to suffer the same academic stigma that Artificial Intelligence was never able to shed.
-
To what "academic" stigma are you referring? Since functional programming underlies many extremely popular and successful computing platforms in academia, including
R(on the FOSS side) and Mathematica (commercial), any such stigma surely hasn't attached to the actual use of functional programming!– whuberCommented Jun 24, 2013 at 17:14 -
@whuber On several occasions I've proposed solutions involving F#. The GIS people shunned it, suggesting that F# is for academics. That may be more of a reflection on the sometimes parochial (spatial-is-special) nature of the GIS community than of the actual technology. It reminded me a lot of the criticism I heard when someone proposed an AI approach to a GIS problem in the early 90's, using Cyc to illustrate that AI is ready for prime time. Commented Jun 24, 2013 at 17:59
-
As flawed as the methodology of this language popularity chart might be, there is discernable rise among functional languages. Commented Dec 15, 2014 at 7:32