Students will work in teams of 3-4 to learn about a useful application, extension, method, or trend in spatial databases, and design a tutorial workshop to present to the class.
The tutorial should be structured as an approximately one hour workshop. The first few minutes can be devoted to an overview of the problem space. What business need are you trying to meet, or what analytical problem are you trying to solve? This may be accompanied by slides, but this is not a requirement, and definitely do not do several minutes of text-heavy slides. Following should be the workshop. This should be hands-on. The students should be given a workbook with clear instructions. You can live demo some of the steps. Code (e.g. SQL queries) should be clearly explained. And your team will circulate and help students with individual challenges.
Topics
The range of possible topics is quite wide. Below are some suggestions:
- Aspects of PostGIS which were not covered in the syllabus, such as:
- Linear referencing
- Temporal support (time tracks, e.g. for transportation or animal movement)
- Raster management and analysis
- PostGIS extensions, such as:
- pgRouting
- PostGIS topology
- TIGER geocoder
- Other applications that support PostGIS
- Various ETL applications, such as GeoKettle
- Integrating PostGIS with R or Python
- NoSQL spatial databases
- There are several major NoSQL products, including MongoDB, Cassandra, and MonetDB; the focus should be spatial support, and you will have to investigate a candidate NoSQL products to make sure that it even has spatial support, but also that it is mature enough to be reliable (i.e. is not in an experimental or unstable phase of the development cycle)
Workbook
Your workshop should be accompanied by a workbook. The workbook should include things like software installation and data downloads. Depending upon your topic, you might provide students with data already available (on Kropotkin or in a virtual machine), and spend your time in the workshop showing how to analyze or visualize the data, but the workbook should have the complete steps provided so that someone could set up the test data without you providing it for them.
The workbook may include screenshots. Code should be included as text, not as a screenshot of a code editor! The workbook should also include links to documentation or learning resources related to your topic. These should be included in a reference section, but also consider including live links at appropriate points in the instructions.
I recommend that you construct your workbook in Markdown. You should export it to PDF and/or HTML. There are a number of converters that can do this, but a very easy one to use is RStudio, which you will also make use of in GUS 5162 and some of the PSM electives.
The workbook should have your names on it. This should be a document that you would feel comfortable posting publicly as a way to demonstrate your facility with these tools. It is also something that you would consider using to run a public workshop at a user conference or industry event.
Workshop
The workshop itself should be practical, and focused on accomplishing a particular outcome, which may be a data management outcome or an analytical outcome. Walk the students through steps to learn how to accomplish the goal. Examples should be clearly explained. Consider having a small number of additional exercises at the end, for which answers are not provided.
During the workshop itself, you should provide live demos or code walkthroughs. Then, as the students work through the workbook, you should circulate among the class, answering questions and helping troubleshoot. Bring the class back together at the end, perhaps showing the final intended result, or showing a further analysis or procedure that interested students could pursue on their own.
Deliverables
- Slides (if any) or presentation outline. These should be provided to Prof. Hachadoorian approximately one week before the tutorial. Slides can be sent via email (PDF preferred) or shared as Google Slides.
- Preparation Instructions. Send instructions to the class for anything that has to be downloaded or installed at least 48 hours prior to the workshop. Note that if the point of your tutorial is itself a complex software installation (such as a server product), this might not be something you have the class do ahead of time. This really is dependent upon the focus of your tutorial.
- Workbook. This should probably be about 10 pages of text. With screenshots, code listings, and sources, it could easily be over 20. The exact length is a little difficult to predict ahead of time.