School of Science and Technology 科技學院
Computing Programmes 電腦學系

A Novel Visualization System for Analyzing the Performance of Legislative Council Members

KWOK Chi Keung

ProgrammeBachelor of Science with Honours in Computing
SupervisorDr. Andrew Lui
AreasData Visualization
Year of Completion2011


The aim of the project is to design a visualization program helping public to analysis the performance of Legislative Councilors from official minutes. We define assessing criteria based on the behavior of members during meetings for example, suggesting idea, interrogating to government, voting and debating of the motion and reporting the voice of public. For instance, a good member should represent general public in meetings and monitoring government. They also suggest idea or ask some questions to improve government’s motion to suit the public needs. In other words, they should attend all meeting and speak in each meeting.

Each minute record the action of member in the Legislative Council meeting but the content is too complex and long, public feel difficult and boring when reading these minutes. We use minutes as our data source and apply text mining technique for processing. Visualization can reorganize all information after text mining processing into graphical form of representation which make public read the minutes faster and easier. It also help public visualizes the performance of a legislative councilor.

The project objectives are described in detail below:

  • Collecting all minutes from 2004 to 2009 and change all minutes including meeting, panels, committees, Subcommittees from PDF format into text file.
  • Identifying the performance attributes of each Legislative Councilor by text mining techniques. Attributes are name, speech content, speech topic.
  • Classifying all topic occur in the minute into 14 areas.
  • Classifying member content by using SOM.
  • Extracting keyword from each member speech content. It uses word co-occurrence statistical information.
  • Statistical techniques apply in the system to count data to form all indicators.
  • Building a visualization application program represent result in part 2-6. It uses jtreeMap, ability graph, SOM and keyword box.
  • Build up a web server for storing the output of part 3 &4. Public can get information by visiting the website and doing evaluation questions.
  • Evaluation of the system: Public can evaluate the rating for each member in the Legislative Council base on the performance in the meeting first. It also asks public rating the indicator using in system and other factor can be determent out of this system and asks public this program is or is not increase the knows of LC member. It also compares the different of the order of performance of member before or after using this system.

Background and Methodology

The election for the fourth term of the Legislative Council of the HKSAR was held on 7 September 2008. According to the Basic Law and the Legislative Council Ordinance, the term of office of the fourth term of the Legislative Council of the HKSAR is four years and started from 1 October 2008. The Legislative Council of the HKSAR has 60 Members for the fourth term, with 30 Members returned by geographical constituencies through direct elections, and 30 Members by functional constituencies. So, each voter at least chooses four to five people form ten or more lists of candidates in each electron. Many voters do not know the candidates well. The performance of proposed Legislative Councilor is main factor of the voting. Voters may feel bushed to analysis the performance because the meeting time is too long and complex and large number of meeting were held during the legislative term.

By using this system, public can easily differentiate the hard working legislative councilor, the member who is enthusiastically bringing ours voice into Legislative Council and supervise the HKSAR government by voting. Public can choose the right person to represent them in next Legislative Council Elections from the visualization of the minutes. It can solve follow problem: in each meeting, motions are included like, the tabling of subsidiary legislation and other papers; reports and addresses; statements; questions; the scrutiny of bills; and motion debates. Some important motions may occur in two or more meetings. It makes the public difficult and boring to trade the action of the member form a long minute. System can find the factor of determent performance of LC member automatically.

Each member can know their performance to self-criticism which area do is not enough in end of each year.

The purpose of Text Mining is to process unstructured information (textual), extract meaningful numeric indices from the text and make the information contained in the text accessible to the various data mining (statistical and machine learning) algorithms. Information can be extracted to derive summaries for the words contained in the documents or to compute summaries for the documents based on the words contained in them.

The key text mining techniques that were applied in this project include text categorization (or text classification), “bag of words” (BOW) representation, self-organizing maps (SOMs), document visualization, performance visualization efforts and treemap. The summary of the techniques and methodologies is shown below:

In order to offer evenhanded standard with a scientific way to study, I have reviewed the participation and commitment of the member through the following major criteria: 1. Meeting attendance rate; 2. Policy formulation; 3. Bill vetting; 4. Motion debates and 5.questions. The details are shown in the following:

The details of the system flow chart is shown below:

All statistics of participation used in this report was obtained from the official website of the Legislative Council Secretariat from 2004 to 2008. All materials stem from and end in the official webpage of Legislative Council on Dec 31, 2010. The target of the program are voter, public and member. The main interface of the system is shown below:

By treeMap visualization, we can easy to know which person performed better. The following shows the ability graphs which mark in the ability come from the indicators in five dimensions. The result sharply indicates Lee Kwok Bo is a lazy member compare to Ho Chun Yan.

Treemap also provide two sorting methods, they are group by the party and sorted by the weight of the legislative councilor. The first sorting method make user easy to compare performance of party by the party block. Second sorting method is sorting legislative councilor block in the party block. The largest one place on the right top one, the smallest on is put on the left bottom of the party block. User quickly knows the influent of legislative councilor.

Other than the treemap, classification of topic use keyword matching method. It only uses the topic header to classify all content. It may have error if a member says something which is not related to the topic. In fact, chairman keep checks the member’s speech is or is not related to topic, if it is not related, chairman will remind the member back to the topic. So the member speech content should relate to topic most of the time, only a little part of speech may off topic. In the program, I assume that all content is related. I use first three year topic list as training set and use fourth year topic for test case. The recall of test case is 89 %, the precision of test case is 100%. As new keyword occur in last year, so recall is not 100%.

Many topics are related to legislation as the main work of LC is enacting bill. The following graph shows Li Kwok Bo and Ho Chun Yan concerned distributed area. Li Kwok Bo only servers two areas but the act of legislative councilor is not server for only two areas, should be all rounded server for public. The chart show Li Kwok Bo is lazy councilor. The chat of Ho Chun Yan shows each area is exited. He is hard working councilor by the chart. It is also matched to the fact. Overall, the chart shows which councilor performs better.

The following graph shows the SOM and Keyword box. Keyword box shows the key issues which have been discussed and the SOM graph finds out which member’s content is mostly likewise.


A survey was carried out in form of a questionnaire on the Internet between 15th April and 5th May . At the end, 20 people have participated the questionnaire. The aim of the questionnaire is to ask for the user of my program what they feel the system and whether the system can help them to analysis the performance of LC member.

Question 1 and 10 ask reader to list out the best three and worst three members after and before using the system. 80% of people said they do not know LC member very well, so question 1 are not finished all. Most of them choose Li Kwok Bo and FOK Tsun-ting as the worst two members. It matches to my system. Other 20% people answer question 10, the result is different to question 1. It shows the system provide other side considering factor to people.

Furthermore, question 3 asks reader to rate the important assessing criteria. They almost think attendance and number of speech is more important than other. Question 4 asks reader to suggest any criteria is not mention in the system and it can be used in analysis the performance. 50 % of them want the system add the motion or bill voting result. It is quite difficult to do that. The reason is some voting result is not record in the minute. Question 5 ask reader how long does it takes to determine the performance of a legislative councilor. Most of them use 20 minutes to determinant. Question 6 ask reader if the GUI of the system is it clear. 90% of them think the system is clear to user. Question 8 ask which part of GUI present well the performance. Many of people think the achievements graph mostly represent the performance of member. Question 9 ask reader can increase the knowledge of member by the system. All people think this system can increase the knowledge of member.

Conclusion and Future Development

All objectives have been successfully achieved in this system. The program can analysis the performance of Legislative Councilors from official minutes. System provides other evenhanded face to public compare which member has best performance. Public can base on the system to make a decision in Legislative council electron. The visualization part of this system can be extended to other performance analyzing problem such as student exam. Teacher can know student’s weak area and adjust the teaching approaches.

Overall, this system provides a new channel for public to analysis the performance of Legislative Councilor. Public can know more about members and make better decision in the upcoming legislative election. The system can apply to analysis the performance of other council member for example USA congress. System just need turn the indicators formation and redefine the structure of minute. For future expansion of the program, system can update the information automatically in monthly or yearly basis. It will add the comparison between voting platform and the related performance. It is an important point to show the relativity of member. It may add the number of violation of Legislative Council Ordinance and stopping meeting undergoing. The system also can add detailed performance visualization between parties.

Copyright Kwok Chi Keung and Andrew Lui 2011

Jonathan Chiu
Marketing Director
3DP Technology Limited

Jonathan handles all external affairs include business development, patents write up and public relations. He is frequently interviewed by media and is considered a pioneer in 3D printing products.

Krutz Cheuk
Biomedical Engineer
Hong Kong Sanatorium & Hospital

After graduating from OUHK, Krutz obtained an M.Sc. in Engineering Management from CityU. He is now completing his second master degree, M.Sc. in Biomedical Engineering, at CUHK. Krutz has a wide range of working experience. He has been with Siemens, VTech, and PCCW.

Hugo Leung
Software and Hardware Engineer
Innovation Team Company Limited

Hugo Leung Wai-yin, who graduated from his four-year programme in 2015, won the Best Paper Award for his ‘intelligent pill-dispenser’ design at the Institute of Electrical and Electronics Engineering’s International Conference on Consumer Electronics – China 2015.

The pill-dispenser alerts patients via sound and LED flashes to pre-set dosage and time intervals. Unlike units currently on the market, Hugo’s design connects to any mobile phone globally. In explaining how it works, he said: ‘There are three layers in the portable pillbox. The lowest level is a controller with various devices which can be connected to mobile phones in remote locations. Patients are alerted by a sound alarm and flashes. Should they fail to follow their prescribed regime, data can be sent via SMS to relatives and friends for follow up.’ The pill-dispenser has four medicine slots, plus a back-up with a LED alert, topped by a 500ml water bottle. It took Hugo three months of research and coding to complete his design, but he feels it was worth all his time and effort.

Hugo’s public examination results were disappointing and he was at a loss about his future before enrolling at the OUHK, which he now realizes was a major turning point in his life. He is grateful for the OUHK’s learning environment, its industry links and the positive guidance and encouragement from his teachers. The University is now exploring the commercial potential of his design with a pharmaceutical company. He hopes that this will benefit the elderly and chronically ill, as well as the society at large.

Soon after completing his studies, Hugo joined an automation technology company as an assistant engineer. He is responsible for the design and development of automation devices. The target is to minimize human labor and increase the quality of products. He is developing products which are used in various sections, including healthcare, manufacturing and consumer electronics.

Course CodeTitleCredits
 COMP S321FAdvanced Database and Data Warehousing5
 COMP S333FAdvanced Programming and AI Algorithms5
 COMP S351FSoftware Project Management5
 COMP S362FConcurrent and Network Programming5
 COMP S363FDistributed Systems and Parallel Computing5
 COMP S382FData Mining and Analytics5
 COMP S390FCreative Programming for Games5
 COMP S492FMachine Learning5
 ELEC S305FComputer Networking5
 ELEC S348FIOT Security5
 ELEC S371FDigital Forensics5
 ELEC S431FBlockchain Technologies5
 ELEC S425FComputer and Network Security5
 Course CodeTitleCredits
 ELEC S201FBasic Electronics5
 IT S290FHuman Computer Interaction & User Experience Design5
 STAT S251FStatistical Data Analysis5
 Course CodeTitleCredits
 COMPS333FAdvanced Programming and AI Algorithms5
 COMPS362FConcurrent and Network Programming5
 COMPS363FDistributed Systems and Parallel Computing5
 COMPS380FWeb Applications: Design and Development5
 COMPS381FServer-side Technologies and Cloud Computing5
 COMPS382FData Mining and Analytics5
 COMPS390FCreative Programming for Games5
 COMPS413FApplication Design and Development for Mobile Devices5
 COMPS492FMachine Learning5
 ELECS305FComputer Networking5
 ELECS363FAdvanced Computer Design5
 ELECS425FComputer and Network Security5