You are on page 1of 15

PLANNING AUTOMATIONS

Optimalisasi Ketersediaan Data Spasial dan Pemanfaatan


Teknologi Artificial Intelligence dalam proses penyusunan RDTR

Jakarta, 11 December 2023

Jati Pratomo

Prepared for: FGD Mewujudkan RDTR Berkualitas


untuk mendorong Pertumbuhan Perkotaan yang
Berkelanjutan Guna Mencapai Indonesia Emas 2045

1
AI, Machine Learning and Deep Learning
Artificial Intelligence, Machine Learning, and Deep Learning have become the most talked-about technologies in today’s commercial world as
companies are using these innovations to build intelligent machines and applications. And although these terms are dominating business
dialogues all over the world, many people have difficulty differentiating between them.

Smart system and machines that can carry out tasks that typically require human
intelligence. It includes the process of imparting data, information, and human
Artificial Intelligence intelligence to machines. The main goal of Artificial Intelligence is to develop
self-reliant machines that can think and act like humans.

Machine learning at its most basic is the practice of using algorithms to parse data,
learn from it, and then make a determination or prediction about something in the
Machine Learning world. So rather than hand-coding software routines with a specific set of
instructions to accomplish a particular task, the machine is “trained” using large
amounts of data and algorithms that give it the ability to learn how to perform the
task.

Is a subset of machine learning that uses artificial neural networks to mimic the
Deep Learning learning process of the human brain. Uses on artificial neural network to reach
accurate conclusion without human intervention

2
Evolution of AI
The term “Artificial Intelligence” was first coined by the father of AI, John McCarthy in 1956. But the revolution of AI consequently began a few
years in advance, i.e. the 1940s. Now, the motivations is to achieve Artificial General Intelligence (AGI)

Artificial Narrow Intelligence

● 2020: ChatGPT 3.0


● 2023: ChatGPT 4.0
● …
● …
● Artificial General
Intelligence
● …
● …
● Artificial Super
Intelligence

3
Challenges: knowing how to utilize
Several factors contribute to the relatively slow adoption of AI in various industries and sectors. Overcoming these challenges often involves a
combination of education, addressing ethical and security concerns, strategic planning, regulatory developments, and building a supportive
organizational culture.

Some common reasons include:


● Lack of Understanding of complexity in implementing AI technologies
among decision-makers regarding how AI works and its potential
benefits.
● Implementing AI systems can be expensive, including costs associated
with software, hardware, training, and integration with existing systems.
● Concerns about the ethical use of AI, data privacy, and potential bias in
algorithms can lead to hesitation and resistance.
● The fear of data breaches and the misuse of sensitive information can
be a barrier to adopting AI, particularly in industries dealing with
confidential data.
● Rapid advancements in AI technology often outpace the development of
regulations, creating uncertainty about compliance and legal
implications.
● Shortage of professionals with expertise in AI development,
implementation, and management, making it difficult for organizations to
effectively adopt and utilize AI technologies.
● Organizations may be cautious if there are limited successful use cases
or demonstrations showcasing the practical benefits of AI in their
specific industry.
4
Challenges: data availability and quality
To produce a good spatial planning product comes for a good analysis, which also need a good quality of data. Nonetheless, the effort to collect a
good quality data often viewed as “spending”, instead of “investment”.

Data silos within the agency Consolidations Compilations Analysis & Modeling
Applications built for cities are often incompatible with one another.
Yet cities need to select the best applications for a specific purpose.
They are then responsible in-house to enable data to move between
the systems, or bridge the systems, to remove data silos and
eliminate the need for staff to manually duplicate data entry.

Data sharing
Local governments need to share & receive data from other agencies,
local businesses, and the public. However, applications aren’t
compatible with non-native data formats, and data is often structured
in a way that works best for the originating system and organization.

5
Data Pipelines
The effort to create a scalable and reliable data pipelines to ensure the data driven decision making

Collection
Data sources (mobile apps, websites, web apps, microservices, IoT
devices, etc.) are instrumented to collect relevant data.

Ingestion
The instrumented sources pump the data into various inlet points
(HTTP, MQTT, message queue, etc.).

Preparation
It is the extract, transform, load (ETL) operation to cleanse, conform,
shape, transform, and catalog the data blobs and streams in the data
lake; making the data ready to consume.
Designing data pipelines architecture is a critical task in ensuring
data driven decision making. It also ensure a reliable and scalable Computation
process within entire data utilization process. This is where analytics, data science, and machine learning happen.

However, on top of everything, in transforming into data driven Presentation


decision making, we also need to understand what is our problems, The insights are delivered through dashboards, emails, SMSs, push
and thus it reflect on what information needed, what data need to be notifications, and microservices. The ML model inferences are
collected and what kind of analysis are needed. exposed as microservices.

Collecting data by itself is a spending, unless we are sure that the


data can answer our problems.
Planning Pipelines
Integrating data pipelines concept into detailed spatial planning process

Spatial planning is a cycle, which includes from the


creations of planning document, implementations
and controlling. Thus, it is a continuations of process
where the output of one process is an input for the 3

Pe
n

n
following process.

ge
naa

nd
nca

ali
Data also has their cycle, which started from data

an
re
Pe
collections, ingestions, preparations, analysis and
pr
presenting the result for decision making.
5 pr
2 5 es

ct
en
es
1

lle
ct
ent t

lle

Co
Co
Therefore, to ensure the continuations of planning
Pemanfaatan
process, the idea is to streamline the process in data 4 1 4
1 ETL
life cycle with the planning life cycle. ETL

pute
Ing
pute
Ing

est

com
All data and result then shall be stored in data

est

com
warehouse, where it can be called anytime when pr
5 es
e 2 3

ct
needed. The transformations of each process is 2 3 nt Prepare

lle
s
Prepare

Co
done through automating the Extract Transform
Load (ETL) process to ensure the consistency of the 1 4
process and increase the resources’ efficiency. ETL

pute
Ing

com
est
Data Lake
Data Warehouse 2 Prepare
3
How AI may help Urban Planning and Management?

AI can significantly contribute to urban planning by providing valuable insights, optimizing processes, and enhancing decision-making. It may also
enhance efficiency, improve resource allocation, and contribute to the development of more sustainable, resilient, and livable cities

Data Analysis and Predictive Modelling Risk Assessment and Disaster Management
● Traffic Management ● Predicting natural disaster
● Simulating Population Growth ● Emergency response optimizations

Smart Infrastructure Land Development and Planning


● Electric grid management ● Zoning optimizations
● Water management ● Suitability analysis for Real estate development

Urban Mobility Public Engagement and Participation


● Public transit optimizations ● Sentiment analysis regarding public
● Ride sharing services perceptions

Environmental Sustainability Sustainability and Resilience


● Air quality monitoring ● Energy efficient building
● Green space planning ● Resilient infrastructure

8
How AI may help Urban Planning and Management?
Case 1: Building Footprint Extractions using deep learning

We implement Generative Adversarial Networks (GANs) for extracting building footprint from Very High Resolution (VHR) imagery. GANs is a deep
learning methods where two neural network compete each other in producing higher accuracy on its predictions result. We present the result in FIG
Conference 2021.

Input: VHR Imagery with 40 cm Training data by making binary All the processes are automated by Common
resolutions map: red is building, while black geospatial libraries GDAL and OGR. 50 training
is non building. sample tiles in total. Train model – 150 epochs
used.

9
How AI may help Urban Planning and Management?
Case 1: Building Footprint Extractions using deep learning (Result)

We tested the result using the image produce in the same area and tested when it implemented in the different area. OA is 82%, which give the
hope that this model having the potential to be scaled up in the different area.

Using the same training data, we implement the algorithm in


Subset from VHR with Training data using Output from GANs different area, and we able to produce the result with 82% 10
40 cm resolutions binary map. accuracy.
How AI may help Urban Planning and Management?
Case 2: Predicting Future Land Use Using Cellular Automata

A Markov cellular automaton (MCA) is a type of cellular automaton that incorporates probabilistic or stochastic rules in its evolution. It consist of a
grid of cells, each in one of a finite number of states, and evolve over discrete time steps based on a set of rules that determine the state of each
cell in the next iteration. Common applications of MCA is to predict future land use by understanding current land use changes.

2020 2030 2040


11
How AI may help Urban Planning and Management?
Case 3: Feature Extractions using panoramic imagery and computer vision

AI can significantly contribute to urban planning by providing valuable insights, optimizing processes, and enhancing decision-making. It may also
enhance efficiency, improve resource allocation, and contribute to the development of more sustainable, resilient, and livable cities

Survey is conducted using 360 camera Calibrated images give the Using deep learning analysis, we able to extract
attached in the car roof. The camera is opportunity to combine with GIS information from images, such as: utility pole and building
equipped with high accuracy GPS and data, direct distance measurement classifications (commercial and non commercial)
IMU sensor. In one day, we manage to from 360 imagery, registering
cover 40 km length of road. In a case spatial information (e.g., point of
of the survey need to be done in an interest) from the images and
area where car cannot pass, then we exported to GIS ready data.
move the camera into backpack.
12
How AI may help Urban Planning and Management?
Case 4: Public Sentiment Analysis

AI, particularly natural language processing (NLP) techniques, plays a significant role in sentiment analysis. Sentiment analysis, also known as
opinion mining, is the process of determining the sentiment expressed in a piece of text, whether it's positive, negative, or neutral. In this slide, we
show the example of analysis using keyword: Penanganan Banjir Jakarta for the entire durations of 2023.

Engagement Across 2023 Sentiment Analysis 2023

Heru Budi continue


building giant sea
Sodetan Ciliwung is wall after warned
inaugurated by Jokowi

Jokowi criticize the river


normalizations that on hold Sodetan Ciliwung is
during Anis period inaugurated

Number of article regarding penanganan banjir jakarta showing the peak in


24 reactions were collected, of that amount 67%
January, February and July. However, although the number of article is high in
was Haha, then followed by love (37%)
January and February, the engagement rate is not high. Meanwhile, the
engagement rate and number of articles increase significantly on July 2023. 13
How AI may help Urban Planning and Management?
Case 4: Public Sentiment Analysis (2)

Social media and online media is a very useful tools to


understand how public react on specific issues (public
sentiment). It gives an insight for policy evaluations,
including gathering public perceptions of urban related
issues.

14
Thank You

15

You might also like