This file contains a set of ideas that, with a bit of refinement, could make reasonable MA, or perhaps even PhD, topics. I have assembled them here in part in hopes that someone might be enticed to work on one of them, or at least see that there are more interesting topics to investigate than any of us will get to in a lifetime.
As a graduate student, for the discipline's and your own sake, it would be better for you to identify a topic that is deemed important by your intellectual colleagues, rather than one that happened to be simply "interesting" or even "funded." Finding a topic that is all three both is ideal.
Articles and books that catalog research priorities offer a budding researcher good advice on topics. Articles with titles that include phrases like: "Grand Challenges" and "New Research Directions" are often such repositories. Cartographica (vol 30 no 4, Winter 1993) included an ICA report, "The Selected Main Theoretical Issues Facing Cartography". Most AAG Presidential Addresses contain pointers. ESRI's ArcUser magazine (in Fall 2010) ran an article "GISicence and Grand Challenges" by Michael Gould that includes several perspectives on research needs. A 2010 book from the NGA, New Research Directions for the National Geospatial-Intelligence Agency , is another case. The NSF Directorate for Social, Behavioural, & Economic Sciences (SBE) has released Rebuilding the Mosaic (NSF 11-086) (look here), which might shed some light on the research environment of the near future. Read the literature for more. But, decide for yourself.
On-line commercial databases such as Google Earth's may lack authority and standing but raise ire anyway. Several recent conflicts (Nicaragua in Nov 2010, Germany in Feb 2011) got great attention, but there probably are other places where the choice of border representations is in dispute. A number of questions suggest them selves.
The USGS DLG (digital line graph) file specification included a generally un-used capability of handling a text field of variable length. This might have been intended to carry feature names, but it might have also been used to carry any other data, perhaps as "tag - value" pairs, or even as free text. Why was it included? Did anyone ever use it? Why not? It seems so darned rich -- presaging "object-oriented" GIS by quite a bit, and avoiding some percieved limitations of the relational model for GIS!!! (Read USGS Circular 895C and the DLG Data Users Guide. Talking with Warren Schmidt or Atef Ellasal might be the quick entre. Keith Clarke and John Cloud might have run across pointers in their research on Corona.)
How about designing a GIS package that treats everything as if it were on the surface of a sphere, spheroid, geoide, or even a detailed terrain model of the Earth, instead of on a flat map? What would a demonstration GIS that treats lines between points as great circles rather than projected straight lines look like? Extra-cool to do it as a cloud mashup with existing web-distributed data. Are interface issues that differ from plane GIS?
We have one thesis underway in this area using GPS-derived locations. There is room for more studies and other approaches. For instance, suppose one takes USGS control points for the islands (now in DXF format, thanks to Ev, Julie, and Jeff) and throws them against the (various sets of) DEM data by interpolating the elevation at the control point's planimetric location. What will the errors look like? From where do they derive?
A good literature review and some experiments to optimize map design for smart phones would be interesting. It might examine particular applications and use scenarios. (See the earlier "Mapping on a Gameboy" below).
Observe people learning the software and note the areas that cause them difficulty. See how this is related to their background and whether the difficulties can be predicted (and thus trained around) and/or avoided by a modified (improved or customized) interface design. One idea would be to come up with objective ways of assessing how well designed (how learn-able) a particular piece of software is.
The study should examine the changing patterns of neighborhood services such as restaurants and small grocers as rents have bid these groups out of many locations. Cases in point are Kings Bakery/Diner on King Street, Yama's fish market, Larry's auto-parts and the TV repair place forced out of Manoa Market place by management non-renewal of leases to accommodate a new Band of Hawaii, the family Chinese restaurant in the Market City Plaza at the corner of King and Kapahulu rent tripling and a new catering coming in after they were forced out, the Kewalo Restaurant forced out by the Harbors Division, and the continued threat of Nick's Fishmarket of Fishermans wharf at Kewalo Basin closing. The gambit may be being used to keep the Tahitian Lanai open. The change of the tire repair shop at the corner of St Louis Heights and Waialai to retail space, even when there is lots of vacant retail space just up the hill in Kaimuki is another odd artifact of this transition. It seems that land rent is forcing many neighborhood functions to close. How does this impact quality of life? Does it reflect or perhaps force a change in the economic composition of an area?
When we were in Miloli'i in Jan 92 we saw a desalination plant that was being built by the community on Homelands at Miloli'i. It provided potable water, brackish water for charging a fire hydrant system (allowing mortgages on house building projects) and other household uses, and also provides a large cold storage room, which fisherman use to keep the catch fresh before being taken to market (in Kona or Hilo). Looking at the community independence marked and facilitated by this project would be of interest. How does it impact the way of life here? What prospects are there for similar projects elsewhere? What features of the sight and community are responsible for the degree of success?
Investigate the issues involved in building GIS/mapping system that could be ported to a Gameboy or similar appliance computer to provide portable information about a city or region. This could presuppose a particular type of use (tourist, home buyer, etc). It could focus on memory resource-sparse symbology and data structures or on the information needed for a particular use. Building a prototype on a PC would be great, but putting it on the actual game machine is possible. Comparing alternative designs would be even better.
Have you ever been in a plane, noticed some feature on the ground, and wondered what it was? Then this one's for you. Design and prototype an electronic atlas/GIS for air travelers with which they could ascertain the names of features that they spot. It would require a database of geographic/geologic features, perhaps a little information about airport locations and air routes. You might be able to assume that planes stay within some bounds between airports, fly at some approximate altitude, that the view would extend out to some distance beside the plane, that these parameters limit the sizes and locations of features that would need to be included in the database. Perhaps approximate flying time along the route could be used to aid database search. It should run on a laptop, or a Gameboy-like thing. [It's funny how electronic books change the landscape but the topic remains open.] It might be marketed to the airlines.
Are there better ways to present users with GIS functionality? That is, can you make it easier to learn, quicker to accomplish tasks, less error prone, easier to refine an analysis, or better by some other measured by criteria)? Build and evaluate alternate (icon-based?) interfaces for existing GIS software (e.g., IDRISI, GRASS, Odyssey, ArcGIS, QGIS, or even a collection of GIS/mapping functions). These could be presumed to be for a specific application or a general interface to the full functionality. It would be best to compare several alternatives but it might be possible to base a design on models in the literature and then see if the performance predicted in the models is realized by the design.
The purpose is not to build a GIS so much as to evaluate alternative interface designs for it --- to examine how people interact with it. (For instance, envisioning a screen in which the user clicks on data icons and on function icons to produce a command that the underlying GIS executes, given the screen described above, do people click the action or the data set first? Do they mind the system going ahead when it has sufficient information to specify a command or do they want to explicitly say "go"? and myriad other questions.)
Try to build an interface to GIS functionality using Pie Menus. Don Hopkins has been using these for various things for 10 years. They make better use of the mouse pointer and movement, working relative to the pointer's position and producing subsequent pie menus in the same relative location fashion, thus not requiring one to point to a target and pull it down or click, and at the same time allowing the expert to mouse ahead in menu choices faster than they can be displayed. Hide MapObjects, IDRISI or ArcView/ArcInfo functionality behind this and do a user evaluation for learning, ease of use, and/or satisfaction. This one would be really cool.
What about building a tool that could take a data table, recognize the kinds of data in in column, prioritize and map them into visual variables, and render the data as VRML? (Rereading Gibson while doing this might be in order, as would making it very net-centric.) Another Java and distributed data project perhaps.
There is considerable recent interest in mapping data quality and or data uncertainty. A good examination of what works and what does not work would be an invaluable step in clearing the largely unsubstantiated set of opinions about how to show uncertainty. Before that, a thorough treatment of the mental calculus used in incorporating information about the certainty of information would be illuminating. The recent literature seems to ignore the (forgive the spellings) Khanaman and Tversky work on people's knowledge of prior probabilities influencing their interpretation/decisions regarding information is now more than ten years old but seems not to be considered by the GIS crowd. And it has probably been refined. How (if in deed at all) do people use uncertainty information in using other information?
(Google Earth and sketch-up have over-run this one.) Assess VRML and worlds built in it as replacements for more traditional maps. Does one learn better from an interactive slightly immersive environment than from a map? How well do people deal with the types of symbology that is easily rendered in available VRML environments? What are the human-factors issues and limits that cartographers will run up against? This stands to get at the real core of what geography (as Earth-writing) can be in the near future.
What do people learn from surface diagrams in various symbologies (2 1/2 D surface diagrams, contour maps, plastic hill shading, illuminated contours, block diagrams, fish-net diagrams, hypsometric mapping, etc.)? Assuming that they learn surface configurations in a deep way, rather than by memorizing patterns of marks of a particular symbol, do some symbols work better than others? Read Malcolm Eley (I have a small collection) and also Kulhavy and Stock in AAAG 86(1).
Calibrate the amount of shift in a map context. What parameters of a map display are the ones that drive the degree of s.c. shift, and how great is their impact(s)? refer to Laura Sewall's dissertation "Determinants of Achromatic Constancy" 1990 - I have a copy from UMI.
Is there still a "local" fleet? Where is the current fleet based and where are its markets? What are the economic impacts on Honolulu and Hawaii? Who makes and supplies the boats? Who buys the fish? Where does it go?
The state's effort to clean the place up has resulted in the closing of the Kewalo Restaurant (was that the name?). This was the only place in the area to service the needs of the people working on the boats in the basin (inexpensive food, grungy dress code, cheap beer, lottery tickets and occasional live music). Since the closing and demolish ion of the bar some (at least one) of the fishing boats from the basin have been sunk. (Check the newspapers for/ around March and April 1993 for one scuttling). Was this a response to the changes in the neighborhood? Was someone taking the best economic route out of the situation in the basin? Are others following suit? What has been the turn over of dock rentals, boat owners, and uses in the basin of late?
What is up with the people living on the hook in Keehi Lagoon? There are ca 50 boats of variable sea-worthiness anchored at State buoys in the lagoon. These boats are under increasing scrutiny as the State clamps down on them. They can no longer stay without being on a State buoy. What is their impact on the local economy? Is this a unique socio-economic group or just people that like boats? What is it like to live in this his community? Why do they choose to live this way?
Hawaii, over the past several years, has moved to limit the number of anchorages and boats in and around Honolulu. In the Keehi Lagoon, the state seems to be going back on a promise that was made when the airport was built that the lagoon would be maintained for recreation and boating, and in place of this are talking about filling in the lagoon to create land for industrial development (personal communication from Mark Smaalders about 10 April 1994). Hawaii is heavily dependent on tourism and could develop this waterfront to be a quaint boat-oriented piece of paradise (sort of like Lahaina), (maybe that is the problem). that would attract people. A little bit of "unsavoriness" would probably help the "real nautical" image of such a place. How are the state's policies toward development of the boat-able waterfront likely to play out in the long run? What are the stated objectives? Are they likely to be met?
There have been several recent episodes where inhabitants of Honolulu were threatened by natural hazards. Two are an anticipated hurricane in the summer of 1993 (July?) and the tsunami warning on Tuesday, 4 October 1994. Neither of these events did in fact cause any damage to life or property but they had great potential to do so. One group which was particularly at risk is the boat-dwellers in Keehi and Ala Wai basins. How did they react in each of these situations? What precautions had they taken? What further precautions did they take after the alert for each of these events? Are their behaviors adequately protecting life and property? Are there policy implications to encourage these or other behaviors? etc. Are these related to being on the hook or in a slip?
One might further consider how these threats and responses compare to those associated with noise from aircraft, pathogens and toxic chemicals in the water, violent crime in the areas, and other dangers (potential and/or perceived) of living on the water in Honolulu.
Calculate a group of census statistics for tracts built by random assignment of contiguous blocks to a new set of census tracts. How likely are the patterns in tract data given the universe of possible tracts? (Esentially became Li Zou's dissertation.)
Maps are often used as a way to store and present information. Information may be useful in resolving disputes, especially where the disputants can agree on the information and get down to dealing with how to use it for problem resolution. Assemble a set of case studies where maps and or GIS have been used to support conflict resolution. Evaluate what cartographic practices contribute to or inhibit a map helping resolve a dispute. Does information help resolve a dispute? Under what circumstances?
Survey several populations to see 1) what they believe about the information held about them and 2) what they believe should be fair information practice. This could be done in part to inform people (perhaps by using questions that raise different pieces of information and makes the ability to link them plain). The survey could compare different socio-economic groups, different educational backgrounds, etc. It might be hypothesized that information workers are more aware and more accepting of information being held about them. Anecdotal information about experiences people have had with information holders might be of interest as a side chapter.
What telecom infrastructure connects Hawaii to the rest of the world? What is the internal infrastructure like? How does it influence Honolulu's ability to do/attract business? Where are the resources? What does that do to influence the location of business? How much traffic is going where?
Contacts include Dan Weydemyer in UHM Communication Dept. He pointed to Dick Barber (941-3789) Pacific Telecommunications Council, 2454 S Beretania, suite 302 PTC tries to maintain information on the distribution of telecommunications infrastructure in the Pacific Basin.
A study of the spatial pattern of voting in Honolulu through time would be interesting. Do some areas consistently vote Democratic or Republican? Are some wards "swingers"? (How) Have the geo-demographics in wards changed through time? Do the parties target "swing" wards? Can these be identified in historical data? What do the party bosses think about in campaigning --- do they target specific areas perhaps with door to door canvassing or do they figure that the media are reaching everyone?
Several tools could be added to the IDRISI suite. Improved vector processing in particular would be a boon. For instance, a routine which took a two layers of IDRISI polygons and output an intersection layer would be handy, as would a routine which took an IDRISI vector file of "not quite arcs-" lines and built a layer of polygons. A routine to do routing through a network of lines / arcs could be handy. Automating "nice" map production from an IDRISI database would be also. Several of these might require the creation or assumption of an auxiliary data file to record properties of the vectors, and associated mechanisms for fast access to these files. Even a routine that sorted IDRISI vectors within a file would be handy. Routines to better handle nested polygons and "regions" (defined as complex polygons made of disjoint parts) would be handy.
Implementing a set of functions that process IDRISI vector representations of transportation networks would be useful. The ability to solve routing, minimum travel time, traveling salesman, and network allocation problems would be desirable. Another version might be to generate descriptive statistics from a network of IDRISI vectors. Routines to generate connectivity, inter-point distance, and other summary matrices would also be interesting.
There is little vector based analytical GIS software available on the Mac. It should be easy to move even workstation type software to the Mac (subject to the interface) because the 68xxx uses a large flat address space and has a well developed graphic toolbox. One could get the Odyssey/Whirlpool code from Nick Chrisman, or the MOSS software from BLM. Much of this is FORTRAN, but there must be a reasonable compiler for the Mac, or one could convert the code to Pascal or C/C++.
It should be possible to write an AML that might use C routines and system calls to parse the biolocality descriptions and query an ArcInfo database to determine coordinates. This would be a way to replace all of the Lisp and ugly C code that the LEI project produced and position us to be better friends of Jack.
Use IDRISI or any other GIS to model the applicability of specific placenames to areas on the ground. Look at the regions of overlap and see if you can predict which name will be used by an individual in a specific context. This type of tool would help make a GIS more accessible and is needed for a natural language front end.
On the Big Island, when we went up and down the coast looking at 3-board canoes, we notice considerable variation. It would be interesting to learn whether this was the result of conscious adaptation to variations in typical sea conditions, some random variation, or due to some other process such as who one learned to build from or where one used to live.
Write a program that can display a scanned stereo-pair on screen and allow someone to drive a cursor around as a dot in model. This could allow one to digitize contours from ap. It should allow one to align the images and drive the spot from the keyboard. It should work with common image files and allow one to output contours as easily used vectors. Correcting to proper planimetry needs to be worked out. As a proof of design, working with two 256 by 256 pixel images is sufficient. It may work better to join the images into a single image to work on, but this really does not matter. (Brice Groff's project)
What if we define a boundary around a polygon as a linear feature that has variation (in positional accuracy, width, permeability, or some other fuzzing property) along its width? (Perhaps we treat the inside and outside edges separately rather than as mirror images of each other, and perhaps we treat the variation by dynamic segmentation, as a property of segments, or parametrically as a function of length along the boundary.) How, and how efficiently, do we do overlay? How do we interpret the result? Do we need to make overlay a stochastic process? Can we treat all linear features this way?
The purpose would be to experiment with using raster (field) representations to compute the likelihood that a particular name would be applied to a given place. One idea would be to simulate the degree to which a placename is applied to points at various distances, and across various terrain features from a "core" area of applicability. What rules govern the usage of placenames?
Before TV had the coffee-house on "Friends", Honolulu had the now defunct 'Coffee Talk' and even 'The Revolution Cafe'. The trendy coffee house hangouts spread here, perhaps from Seattle on the mainland. And they seem to still be proliferating. What is the story? How many of these places can exist here? Why did someone finally decide to open one? How has the idea spread here? Do entrepenures encourage each other?
A more general study is suggested here than the coffee shop one. Several trendy innovations have recently appeared around town. Coffee houses, a rock climbing gym, beer-making, collecting POGs, and probably a number of other fad activities could be included in a list. Why and how do these occur? Are there differences and/or similarities in the spatial processes that each of these follows? How would you even begin to do this?
Another innovation diffusion idea. Catalog all of the internet service providers. Characterize them as to services, location, market etc. Interview the principals. How did they learn how to provide services? How did they decide to get into the business?
I don't think this one stinks! Walking around Honolulu, one is struck by numerous aromas. How about a project to map them? The scent of bakeries, steakhouses, gas stations, traffic, the ocean, flowers, sewage treatment plants, etc. all combine to build a landscape of the nose. (It might be interesting to work with blind people to see whether these environmental cues are used by them for orientation.)
There is considerable work on depth cues in line drawings. They are often useful tools (E.g., Davis diagrams of landscape evolution and the illustration at the top of the Los Alamos Lab web pages --- http://xxx.lanl.gov/legal/disclaimer.html). How about an analysis of a set of minimalist drawings of terrain to see (1) how they work, (2) whether they do work for everyone, (3) if there are predictors for when they won't work.
Using either grid cells or parcels as the base units, try to build a simulation of the change in landuse on Oahu in this century. How do land suitability and say, financial need compare as drivers in the process?