Management and Mining of Spatio-Temporal Data   Coordinated by Dr Rui Zhang, Updated on 10 Feb 2013

 

Credit: 12.50 (in the form of COMP90005 Advanced Studies in Computing 6B, new version of the previous "Directed studies 6B")
Level: 9 (Graduate/Postgraduate)
Time: 36 hours contact, 9 days, each day includes 3 hours lectures followed by 1 hour lab (the time of lab may vary on a few days).
Prerequisite: Students are expected to already have

(i) taken subjects or the knowledge of algorithms and data structures  and most computer science foundation knowledge,
and (ii) C/C++ or Java programming experience
and (iii) experience in basic use of Unix/Linux system.

In general, either one of the following types of students should be qualified to take this subject.

 (i) a Postgraduate Research student (i.e., RHD students)
or (ii) Masters by Coursework student with Bachelor's degree in a computing related major,

but still you must check with Dr Zhang to confirm (see enrolment instructions).
 
Enrolment Instructions: You MUST  first email Dr Zhang to briefly describe your background and get approval. Once Dr Zhang approves, he will directly email Dave Segan <dsegan@unimelb.edu.au> in the Engineering Student Center to get you enrolled in the subject. You might then follow up with Dave after a few days making sure you are really enrolled. Please note that you must email Dr Zhang by 5 Jan 2013 (the earlier the better in case a discussion is needed). What you should describe in your email to Dr Zhang is as follows:

(i) Your student number, your name, and your student email

and (ii) If you don't have any of the knowledge or experience specified in the prerequisite, what equivalent experience you have or how you are going to make it up (e.g., you may spend a couple of days to learn the basics of Unix/Linux before the subject starts).

and (iii) If you are a postgraduate research student, please describe which year of study you are in and what topic you are working on. If you are a Masters by Coursework student, what's your undergraduate major.

Please use "Enrolment for intense subject: Spatio-temporal data management and mining" as the title of email.

Note: If you want to withdraw this subject after enrolment, the latest date for withdraw is 8 Feb 2013.

Provisional Course schedule (number meaning day number, three weeks in total):

Dates Lectures Assignments
11 Feb Introduction to this subject Introduction to databases and indexing; Spatial queries and indexing; 1 hour guest lecture from Chris Leckie on Sensor Networks. Lab 1 specification

Lab 1 code

 

12 Feb Spatio-temporal Queries and indexing.
13 Feb guest lecture from Rao Kotagiri on HMM, CRF , Kalman Filters and Data fusion; Location-based social networks
14 Feb Skills for understanding advanced research papers, writing top conference/journal papers, and paper reviewing in this area. 

Detailed Guidelines for Paper Review

Data structure proposal assignment released: propose any data structure of your own or a new algorithm on an existing data structure to solve a problem of your own choice. See submission and assessment details at the bottom of this webpage.

Presentation of Reviewed Papers released: see instructions beneath this table.

15 Feb Introduction to Map-reduce framework and large-scale spatial query processing with Map-reduce; 1 hour guest lecture from Raj Buyya on cloud computing. First lab due. Lab 2 setup instructions

Lab 2 assignment

18 Feb Preprocessing on trajectories data;     Queries on trajectories;     Trajectory pattern mining Result of Lab 1 basic tasks
19 Feb guest lecture from Lars Kulik on Privacy of trajectories Destination prediction based on trajectories Link to the Hortonworks Image
20 Feb Rest  
21 Feb Rest Draft data structure proposal due for getting feedback (compulsory)  
22 Feb Rest; Second lab due.  
25 Feb Presentations from students¨ reviewing assigned papers Result of Lab 2 basic tasks   Feedback
26 Feb Reserved for miscellaneous matters, selected topics and general discussions.  
27 Feb Rest  
28 Feb Rest  
Extended to 4 Mar Final Report (including data structure proposal writing and reviews) and lab challenge questions due See details of submission instruction at the bottom of this webpage.

 

Feedback for Research Proposal

Feedback for the research proposal has been emailed to all the students. It contains proposals from all the students. Please note that

1. Some of the proposals are the authors' current research ideas, so you must have agreed to keep them confidential if you read them. You are
only allowed to look at other students' proposals and feedback for study purposes.

2. You must not change your data structure proposal in the final report to mimic any other student's proposal.

3. If you receive a feedback advising you to submit another draft before final report: this is not compulsory, but it will be very important
because we believe you have not explained enough new algorithm or new data structure in the proposal.

4. In the final report, if you need more space to present your proposal, feel free to do so. Up to three pages are allowed for the data structure proposal in the final report.

5. All the proposals have been thoroughly discussed between John, Andy and myself. The feedback is entered by John and Andy. You are welcome to ask John, Andy or myself regarding further advice or feedback.

 

Papers to Review for the Final Report

Please choose either Group A or Group B to review. Each group consists of three papers. You do NOT need to email me about which group you choose since I can tell from the final report. A review on a paper should still contain at least four parts: (i) summary of the paper, (ii) strong points, (iii) weak points, and (iv) detailed comments.

NOTE:

1. This assignment is individual work. You should not discuss with each other regarding the reviews on these papers. If two reports have too similar reviews, I might ask questions to these students regarding details of the review before I give the final mark.

2. Some of these papers are rejected submissions, so you should NOT treat the following papers as published papers and you should NOT cite them in your own research paper (if you ever write any) based on the venue you see in the submissions.

Group A Paper a1
Paper a2
Paper a3
Group B Paper b1
Paper b2
Paper b3

 

 

 

Presentation of Reviewed Papers

Please first form groups where each group has three students. Then email me the paper number that your team wants to review and present. A paper will be assigned to at most two teams. If more than two teams are asking for the same paper, the paper will be assigned to the first two teams in first-come-first-served way. The latter teams will be asked to choose other papers.

All presentations will be given on 25 Feb. Every presentation is 15 minutes long plus 5 minutes questions. Each team member should present approximately 5 minutes. The contents of the presentation consists of: (i) summary of the paper (ii) strong points (iii) weak points (iv) other detailed comments on the paper.

Paper 1 Taken
Paper 2 Taken
Paper 3 available
Paper 4 Taken
Paper 5 Taken
Paper 6 Taken
Paper 7 Taken

 

Order of the presentation

10:30 10:50 11:10 11:30 11:50 12:10 14:30 14:50 15:10 15:30 15:50
Xiwei Wang
Spandana Gella
Jose Mendez Dangon
Goce Ristanoski
Yang Lei
Jiazhen He
Shiqin Chen
Boliang Liu
Yi Lin
 
Simone Romano
Zeyi Wen
Jin Huang
 
Han Wang
Xu Li
Fei Liu
 
Joost Kuckartz
Janaka Seneviratne
Sujatha Rajagopalan
 
Saad Aljubayrin
Boyu Chang
tfyang
 
Miao Kang

Yun Zhou

Hengfeng Li

 

Zhihong LU

Yingzi ZHANG

Yichang ZHANG

 

Cagatay Yucel
Elham Naghi Zadeh Kakhky
Sarah Monazam Erfani
Maryam Fanaeepour
Abdullah AlDwyish
Eman Bin Khunayn
Kushani Perera

 

 

 

Discussion Forum

 

Please join the CS Combined Research Projects Community to use the discussion forum "2013 Summer Intense Subject COMP90005" to discuss research questions and form presentation teams.

1. Login to the LMS.
2. Click on the Communities tab click on self-enrol link inside the What is a Community? module to bring up the list of all self-enrol communities.
3. Select COM_00457 CS Combined Research Projects.
4. Click on the double arrows below the community ID to enrol.

 

Lecture recordings are here.

 

Contact time: 10:30am - 12:30pm;  2:30pm - 4:30pm every day. If you are strongly against such time, please email Dr Zhang. 

 

The current offering will be during 11 Feb 2013 to 1 Mar 2013.

 

Venue of the subject:

 

11 Feb to 20 Feb (1st week to Wed 2nd week) Babel Middle Theatre Lab: Alice Hoy 108 and 109 (open 3pm-6pm)
21 Feb to 22 Feb (Thu to Fri 2nd week) Redmond Barry Lowe Theatre Lab: Alice Hoy 108 and 109 (open 3pm-6pm)
25 Feb to 1 Mar (3rd week) Babel Middle Theatre Lab: Alice Hoy 108 and 109 (open 3pm-6pm)

 

Corequisites: None

 

Recommended Background Knowledge: Data management or data mining, but not necessary

 

Non Allowed Subjects: None

 

Recommended readings: Computing with Spatial Trajectories (Springer 2012), a list of recommended papers from the lecture slides.

 

Overview:

As the proliferation of mobile devices, spatio-temporal data is produced at unprecedented speed. We can mine interesting knowledge from this data for wide ranges of applications such as traffic management, route optimisation, location-based social networks, disaster monitoring and urban planning. Faced with a plethora of emerging and novel applications on spatio-temporal data, what knowledge to mine from the data, how to mine the data and how to manage the huge amount of data are pressing issues. This subject provides an introduction to such topics and provides students with basics to enter more advanced research in this exiting area. This subject also provides students with advanced skills in evaluating the quality of research papers in this area and writing papers for top conferences/journals in this area. Compared to spatial databases which studies static spatial data, this subject has the focus on the temporal nature of the data (typically generated by positioning devices like GPS) and the emerging applications enabled by knowledge discovery from spatio-temporally correlated data.

 

This will be an intense subject offered during the summer. Most of the contents will be taught by Dr Rui Zhang. Prof. Rao Kotagiri, Prof. Raj Buyya, Prof. Chris Leckie and A/Prof. Lars Kulik will give guest lectures on certain topics. Please note that the guest lectures might be subject to change based on the guest lecturers schedule.

 

Assessment:

(i) Two lab assignments worth 30%.

(ii) A data structure proposal worth 10%, about 500 words (approximately one A4 page) for the draft proposal and up to three pages in the final report. You must submit a draft proposal by 11pm on 21 Feb by emailing Jianzhong Qi <jiqi@student.unimelb.edu.au> with the title "COMP90005 proposal ID SURNAME", where ID is your student number and SURNAME is your surname. You will get feedback on your proposal. You will have 3 marks deducted if you do not submit the draft proposal. The draft proposal should be approximately or close to 500 words, but you won't be marked, so presentation and writing may not need to be perfect at this stage. The final proposal will be submitted as part of the final report. It will be marked based on the presentation, writing, correctness and novelty. The proposal must be submitted in word or pdf format.

(iii) Presentation of reviewing one paper worth 15%, done in group of three students.

(iv) Reviews on three papers worth 45%, about 2500 to 4000 words (approximately five to eight A4 pages).

The final report consists of the data structure proposal, the reviews on three assigned papers. The final report is due by 11pm on 4 March by emailing Jianzhong Qi <jiqi@student.unimelb.edu.au> with the title "COMP90005 final ID SURNAME". The report must be submitted in word or pdf format.

To pass the subject, you need to pass the following hurdles:

* Lab assignments:            12 out of 30

* Data structure proposal:   5 out of 10

* Presentation:                   7 out of 15

* Overall:                         50 out of 100

NOTE: all assignments and reports are individual work except for the presentation of reviewed paper.