Activities> Final Project
-
-
Types of Projects:
-
Survey/review
Survey the publications within a specific research topics within database
forums in last 5 to 10 years. Sample topics include topics within spatial
databases such as conceptual modelling of spatial data, indexing and
querying collections of moving objects, vector map compression, spatial
data mining, spatio-temporal databases, mobile and wireless spatial
databases, spatial data warehouses, internet based spatial databases,
semantic web, homeland security, using spatial indexes for content based
retrieval, etc.
-
- Statistical summary of the publication activity by topics and year
is often valuable to get the big-picture of research activities in areas
with vast literature. Projects in this category will collect data about
publications and statistically summarize the publication activity across
different research topics (or validation methodologies used) in spatial
database forums in last 5 to 10 years.
-
- You may find more topics from the call for papers (see Symposium on
Spatial Databases, or ACM Workshop on GIS, UCGIS Research Agenda websites)
from latest conferences on databases.
- Extensive sources for survey papers include ACM Computing Surveys,
IEEE Computer (e.g. Embedded Databases survey in 9/2000 issue), and
Communications of the ACM. A recent issue of IEEE Transactions on Knowledge
and Data Eng. (January 1999) also featured a number of survey papers
on spatial database topics. The format of the survey paper may resemble
those used in survey papers presented in prestigious computer science
journals mentioned above.
-
Research Paper
These papers should analyze a particular problem and propose some
solution. Follow the outlines from the research papers covered in the
course. For example, the paper may have 5 sections:
* Problem Statement, Significance of the problem
* Related Work and Our Contributions
* Proposed Approach
* Validation of listed contributtion (experimental, analytical)
* Conclusions and Future Work
-
-
Development
The research aspects of software demo project may be structured by
addressing the following questions:
* Functional Requirement analysis.
* Non-functional requirement analysis.
* Propose a benchmark for designing and evaluating the software applications.
* Identify the key design issues in designing the software.
* Propose and compare alternative of solutions.
* Develop a prototype to demonstrate your work and write a short paper.
-
Example of Topics
-
-
- Mobile database (e.g. Sybase Anywher, Oracle lite, Synchrologic,
...) are becoming important with the increase in popularity of PDAs
and other mobile computing platforms. Spatial applications (e.g. roadmaps,
point of interest, routing, proximity queries) are among the killer
applications in this domain. Explore the impact of mobile environment
on spatial databases.
-
- Implementation and comparison of at least 3 spatial indexing (R-Tree,
quadtree, HR-tree....).This project should implement the data structures
and design experimental evaluations that consider the generation of
data sets, the set of queries for evaluations, and parameter sensibility
of the data structures. The experimental design should allow us to extract
some conclusiones about the static and dynamic performances and use
of resources (memory).
-
- Consider your first day as a student of the University of Concepción.
Identify the information and querying needs to support the major activities
during that day. (Are some of these geographic in nature?). Consider
a Personal Digital Assitants or PDAs. Compare the characteristics (e.g.
CPU, Storage, Communication) of these devices with those of typical
computers hosting traditional DBMSs. Identify the key design issues
in designing a DBMS on PDAs that support students during the first day
on the campus. Develop a prototype to demonstrate your work and write
a short paper documenting the key findings.
-
- Create spatial visualizations for the collection of publications related
to spatial databases or spatial data mining or other topics. The project
would characterize of the goals for visualizing a collection of documents,
choose as well as compare of visualization modes. Example goals are
to assess the poupularity of topics, interaction among topics, assess
temporal trends, e.g. change in popularity of a topic. Candidate visualizations
include
1. elevation map: hills= cluster centers, height= cardinality
2. Quad-tree space division : leaf = document, non-leaf= topic
3. Subway maps: station= document(s), lines= topic
-
- Extend the idea of threading for the leafs of tree based spatial access
methods (e.g. R-tree) by linking nearest neighbors in some (e.g. 4 or
6) or all directions. Threads may improve the I/O efficiency of range
queries, nearest neighbor queries but increase overhead of updates.
Design new bottom-up algorithms for range queries and nearest neighbor
queries taking advantage of the threads. Also extend the update algorithms
to maintain the threads.
-
- Consider a system that needs to keep track of objects that are moving
in a space. The information of objects are collected by several servers
who are in charge of predefined partitions of the space. Discuss issues
about how to answer queries about where and when an object was/is as
well as how many objects move from one to another region in the space.
Keep in mind indexing and query processing. Identify 4 important issues
and develop some solutions.
-
- XML is reinventing databases in many areas (e.g. query languages)
related to conceptual and logical databases. Specific tag families (e.g.
GML for spatial dataset) of XML are being created for different data
types. XML parsers (e.g. DOM, SAX) are becoming widely available for
manipulating XML data outside a database. However, querying XML datasets
is not scalable since entire dataset has to be parsed and processed
even if a small part of the data is relevant to the query at hand. Explore
incorporating physical database notions (e.g. indexing, clustering,
query processing) to process point, range, join queries on XML datatypes.
Design and perform experiments to evaluate the overheads and usefulness
of such information in XML datasets in context of a useful application.
-
- I will add more topics during next couple of weeks and I am open to
suggestions.