---
tags: JC2066
---
# JC2066 Seminar 1

21-10-2022

# Lecture 8 Reflection
## Where would you place the sensors?

Utility poles
- Just the right amount of density
- ~180 million utility poles around the US
- Capable of capturing gunfires in domestic households

## Any problems with this "Big Data" strategy?

- Privacy
- Citizens may not be comfortable with their daily lives being recorded
- Data leaks
- Hackers may steal and publish data to the public
- Costs
- Costs billions to install
- Installation labor
- Material costs
- Network & server costs
- Storage, etc

## What is missing?

- How would the algorithms be designed?

# Examples that should be filled in


# Professor Alfred T. HO
E-government public policy
back to 2011 public policy department

Helps minorites with subsidies etc with public tax money

public and international affairs

30 years in the US

Was in a housing management conference

Apartment complex use a lot of techs

## Introduction
Understand context in public policy

Recruit into Masters programme
decide to get into public service or public policies or smart cities or want to get a social science perspective PHD

Research papers

Will come back next week

## Definition of Smart City and Smart Government

### Smart city
HK government smart city blueprint

Adaptiing smart city blueprint

>Focusing on innovation and technology

>Strong economy and high quality of living

>Very tech driven

- Innovation and technologies are just the tools

- Smart mobility
- Subway system in new york is extremely bad compared to Hong Kong
- Shang hai has a transportation system on par with Hong kong
- Hong Kong is a pioneer in public transport system
- Sensors in different parts to make sure the flow management is good
- Employing different mode of transportation like buses and trains
- Smart systems behind QR codes, Smart Ids in transport
-
- Smart living
- Broadband 99.6% penetration rate
- Rural areas in the US doesnt even have broadband
- Healthcare system
- Dealing with elderly population needs
- StayHomeSafe
- Smart environment
- Reducing emissions and solid waste
- Dealing with the electricity consumption
- Energy efficiency
- Building information management system
- Sensors to check what is going on
- VR to design
- Smart water meter system
- Escalator remote monitoring system
- Using AI to simulate accidents
- When a cameras see a person trip, alert AI system and send staff there
- Drainage blockage detection
- Flooding sensors
- E-log book
- Sensors to monitor fire risk
- Radiative cooling
- Flexible PV panel(Solar panels)
- Sterilizing Robot
- Simplifying public housing construction using modular integrating construction
- Sensors on safety helmets
- Temperature sensors
- Air quality sensors
- Pollutants / Chemicals
- Location sensors
- Tell you how to run and help u navigate

- Smart people
- Education
- Leveraging talents and abilities of non governmental sectors

- Smart economy
- Digital economy
- E-banking
- FinTech
- Smart tourism
- LawTech
- Promote R&D, re-industrialisation

- Smart government
- Deliever information
- Adpot techonlogies in public service
- Use data analytics to improve service delivery, policy making and public management

>New version Smart city 3.0 is coming

## Big Data and Data-Rich Environment

- Collecting data
- Sensors
- Big data algorithms

Big Data definition

- Volume
- Cell-level data
- Big number
- Thousands at least
- Social science rely on surveying, not big
- Complexity
- Structured and unstructured data
- Audio is and example of unstructured data
- Volatility
- Data change all the time
- Becomes irrelevant
- How long should you keep them?
- US police violence
- They wear body camera
- Thousand of police offiers in los angeles out there for eight hours, five days a week
- Storage problem
- More than 150 zettabytes will need analysis by 2025

Cannot only rely on big data
- Small data is important as well
- Human intelligence and insights still matter

Smart government
- Data tapes long time ago
- Don't have hard disks or floopy disks
- Mainframe computer
- Signal staff in the mainframe computer room to mount the tapes

- Administrative records
- Web info
- Social media content
- communication data
- sensors
- cctv
- mobile phones
- smart meters

## Applications of data analytics in public service and management

- Planning
- Custom service
- Citizen communication
- Resource allocation
- Logistics
- Personnel management
- Procurement
- Performance measurement and evaluation
- Interdepartmental collaboration

Public exams
- Singapore
- Public exam cameras
- AI monitor gesture
- Calculate how likely you are cheating
- Send a teacher there


CS guys take a key role in this diagram

US Chief data officers just clean data, very boring

Key to pick up:
- PEOPLE AND COMMUNICATION SKILLS
- TEAMWORK
- PICKUP NON-CS SKILLS
Need to talk to 6-7 departments to get the data
- Talk to people so that they will surrender the data to you
- Make sure they are comfortable
- Data might be full of errors and they might be uncomfortable
- Political skills
- Story-telling
- Departments refuse to share the data
- Set up policies so that they must share data

Heatmap

Boring graph

Cluster graph

Ways of visualizing data is important
- Audience awareness matters
- Interesting graphs like heatmap to the public
- Boring graphs for business clients


- Use data from birth and death departments
- People use dead people to cheat the system and get $

Combine survey data and all kinds of data together

House market value:
- Home age
- Condition
- Acreage
- Bedrooms
- Bathrooms
- Garage
- Basement
- Foundation (brick, etc against bad weather conditions)
- Style
- Good management
- time to highway (decreases the closer it is)
- distance to offices, hotels and retails
- Green investments like parks
- Distance to library
- People with college degrees

Preventive healthcare using big data

- Economic stability
- wont be stressed
- Put sensors in elderly people
- Debatable Topic!!!!!!!!
- Smart toilet
- Putting sensors in toilet bowls
- Dont want to test all the time
- checking what you produce
- what chemicals are inside
- what bateria are inside like covid
- check whether you are in good health
- Intrusive sensors
- Possible Topic!!!!!!!!!

London has the highest density of CCTV in the world
- CCTVs in intersections etc
- Sent to law enforcement in real time
- Virtual police
- Every single police department in the world is using survelliance cameras
- Kansas, US uses ShotSpotter program with sensors to detect gunfire in neighborhoods
- Send police officers to investigate
- Check cameras nearby


# Mexican drone solution
- Drones drop Mexican city's crime rate by 42%
- Gangeters cannot shoot them down because they are so high up
- Super high resolution cameras
- Track down where cars and criminals go

# Sensor placement:
- Utility poles
- ~180 million utility poles around the US
- Problems of the big data strat - ethics privacy blah blah blah
- Data leaks
- Costs billions to install

## Lessons learned

- Importance of leadership and communication skills
- Look into multi-goal optimization
- Roles of different departments
- How interdepartmental communication works
- Presentation of data
- Irrelevant skills ARE relevant

# Chat archive

From Zhicong Lu to Everyone 08:58 AM
Good morning, everyone. Let's wait for a few more minutes so that more students can join. Thanks!
Me to Zhicong Lu (Direct Message) 09:42 AM
Uncompressed 720 8-bit for 40 hours: 8.68TB x 500 officers
From Levi to Everyone 09:42 AM
resolution * fps * 60 * 60 * 8 * 30 * 500 for a month
From Me to Everyone 09:47 AM
Audio size is missing
From Zhicong Lu to Everyone 09:48 AM
Another student Tak Shing WONG answered: Uncompressed 720 8-bit for 40 hours: 8.68TB x 500 officers
From CORNEJO David Alexander Leynes to Everyone 09:48 AM
don't have
From FOR Lek Shyuen to Everyone 09:48 AM
i don't have
From Zhicong Lu to Everyone 09:56 AM
Mainframe computer: https://en.wikipedia.org/wiki/Mainframe_computer
I talked a little bit about privacy-related topics and surveillance, but not directly about personnel management
From Zhicong Lu to Everyone 10:40 AM
Let's take a 10-minute break. Thanks!
From Tat Kei Alfred HO to Everyone 10:47 AM
What Dr. Lu shares about mainframe computer is too nice and modern.  This was the mainframe I used: https://en.wikipedia.org/wiki/IBM_3090
Also, I strongly encourage you to watch this movie if you have not:  Hidden Figures https://en.wikipedia.org/wiki/Hidden_Figures
From Me to Everyone 11:04 AM
hi
Where would u guys place the sensors
There are an estimated 180 million utility poles around the United States today
maybe we can put them there
From Tung Dawn CHAN to Everyone 11:04 AM
Agree
From TANG Cheuk Iu to Everyone 11:05 AM
agree
From Me to Everyone 11:05 AM
Problems of the big data strat?
ethics privacy blah blah blah
From Tung Dawn CHAN to Everyone 11:06 AM
Can nice approach
Eg Leakage of privacy
From Me to Everyone 11:07 AM
yeah
data leaks
From Tung Dawn CHAN to Everyone 11:07 AM
Any other ideas 🫂🫂
From Me to Everyone 11:07 AM
hmm
From Tung Dawn CHAN to Everyone 11:07 AM
Tho I don’t think we have
From Me to Everyone 11:09 AM
Cost?
installing those sensors everywhere in the US must cost millions or billions
connecting them thru a network etc
installation labor
material costs
From Tung Dawn CHAN to Everyone 11:09 AM
How about we estimate the cost to billion
From Me to Everyone 11:09 AM
alright
From Tat Kei Alfred HO to Everyone 11:10 AM
Please type in your brief answers to my questions in the chat, or present orally
From Iusuf NAUMOV to Everyone 11:12 AM
1)Firstly place cameras at schools, in order to protect children, I think it’s more important. Also it is better to Google, where gunfires commonly happen
From CHEUNG Hung Yam to Everyone 11:12 AM
Room 13
I would place them at where there are not much flow of people such as alleys in order to detect any crimes happened at there because some people may think that there would not be any monitors or sensers and commit crimes.
The problem is that there would be too many data that people would be missing and not noticing.
From SIT Lai Wa to Everyone 11:12 AM
for the question 1, I think the sensors should be place the public places, especially the places with many people, this can low the harm of shot accident, like schools, marketplaces. recently, there are some event of shot accidents happened in schools and causes many young children die.
room 6
From Me to Everyone 11:13 AM
Room 5 Sensor placement:
Utility poles
~180 million utility poles around the US
Problems of the big data strat - ethics privacy blah blah blah
Data leaks
Costs billions to install
From WU Ying Kit to Everyone 11:13 AM
[Room 3]
Q1. The roof of the building. It is helpful for gathering evidence.
From YAN Yachao to Everyone 11:13 AM
room7:
mostly the place that may frequently occur accidents
From RAHMAN Abed Al 57744270 to Everyone 11:14 AM
(Room-10) I believe that we should place cameras at intrusion points, like windows, doors, etc. at a mass level, across most buildings in a city. This will make sure that whoever tries to break in, is well documented in the recordings.
From CHAN Ting Fung to Everyone 11:14 AM
Place audio sensors in libraries to detect unsuppressed gunfire. If someone want to shoot in a library, they would have to install a silencer so that it won't disturb others in the lbrary
From ZOU Huiqi to Everyone 11:14 AM
Room 2
Q1: Fixed locations such as buildings, telephone poles and street lights.
Q2: It is difficult to distinguish whether the sound is from the gun or other sound. The accuracy may be low.
Q3: How to dispatch officers to the scene. It may take a long time for the officer to go there.
From CHO Tsz Yan to Everyone 11:16 AM
Room12
places that they have high crime rate.
From CASTRO Justin Campo to Everyone 11:16 AM
[Room 14] 1.) Place in areas with a large crime rate statistically. 2.) One probable problem is that they may be placed in areas with larger composition people of color. And, there may be possible false alarms 3.) Is it worth to spend so much money and effort just for a smaller benefit than just calling the police?
From MAN Kai Wing to Everyone 11:16 AM
Room 7:
1) We think the sensor may be put in the mall and school.
2) There have two problems: Privacy and Wrongly triggers.
3) They may also install a camera to collect criminal evidence.
From HUI Hiu Wang to Everyone 11:16 AM
the sensor should be placed at the place where people visit frequently such as supermarket or restaurants.
From CORNEJO David Alexander Leynes to Everyone 11:16 AM
Room 11: In secluded areas that are away from public view, such as alleyways. These are areas with fewer possible witnesses to crimes, so it is more important to monitor them.
From Iusuf NAUMOV to Everyone 11:17 AM
Room8:
2)Most common problem is memory limit. Also AI/program can struggle with detecting gunfire/faces, maybe because of low quality of cameras.
From FOR Lek Shyuen to Everyone 11:17 AM
room 12
1)Sensors should be placed on street lamps to detect sounds from a high altitude avoiding errors from reflecting waves.

2)Unable to differentiate the gunfire sound from residents or police so that it may cause a false alarm.

3)There should be an alarm to let people run away from that area if a gunshot happens.
From Yue Yu to Everyone 11:17 AM
Room 7. I would place the sensors in the public places that are quiet and hidden. Most criminal behaviours does not happen in the crowd. However, in this case the privacy cannot be guaranteed.
From WONG Kin Nam to Everyone 11:17 AM
Room 4 : We think the sensors should be placed in those areas with high population density. Some false evaluations may arise if the algorithm is not well-designed, like misinterpreting the environment sounds. We may miss some proper legislations to restrict the usage of Big data by the law enforcement agencies
From LO Chung Yin to Everyone 11:18 AM
Group 12

Where to place
Sensors should be placed on street lamps to detect sounds from a high altitude avoiding errors from reflecting waves;
somewhere with high crime rate;
somewhere people in need of protection are frequented, such as next to school.

Problem:
Unable to differentiate the gunfire sound is from residents or police so it may cause false alarm.

A more specific identifier of the gun user 

Can’t detect the sound of gun with a muffler installed
From HUANG Shing Yin to Everyone 11:18 AM
I would put the sensors on the lamp post and school 
because these two place is the most frequency for gunfire.
According to shotspotter.com, the With the technology 
sensitive enough to differentiate 
gunfire from fireworks, explosions and automobile 
backfiring, ShotSpotter eliminates false reports of 
gun shots and accurately places occurances for police 
to respond efficiently within minutes of an event.
Therefore i think the big problem isthat some error or 
misjudgment will be happened.For example, the sensors may
detect gunfire when citizens only let off the fireworks.
Maybe we missed the time of police arrive the criminal
place.
From CHAN Ka Hong to Everyone 11:18 AM
1. Likely some areas near schools where most recent shootings occured
2. maybe there are false alarms e.g. sounds similar to gun shots -> possible incorrect data collected
3. maybe missing out places where there are no sensors; cannot detect non-gun shots
From Me to Everyone 11:33 AM
chase down criminals
From CHAN Ting Fung to Everyone 11:33 AM
drones can drop grenade
From CORNEJO David Alexander Leynes to Everyone 11:33 AM
it acts as a deterrant?
From CHAN Ka Hong to Everyone 11:33 AM
capture their faces?
From HO Ping Hong to Everyone 11:33 AM
perhaps recognize their faces?
From HENDRA Latisha Besariani to Everyone 11:33 AM
capture criminal acts?
From CORNEJO David Alexander Leynes to Everyone 11:33 AM
I think the fear of being caught is enough to cause fear
of being caught
From RAHMAN Abed Al 57744270 to Everyone 11:33 AM
Drones can follow their tracks, to locate them accurately and deliver the information to the police.
From above
From WU Ying Kit to Everyone 11:33 AM
avoid police get killed
From FOR Lek Shyuen to Everyone 11:37 AM
when too many people gather, point red light on them
but the victim has already passed away