02.12.2020 Views

Urban Sprawl Poster

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

URBAN SPRAWL

Urban growth and urban sprawl are sometimes used synonymously by the people, although they are different. Urban growth is a sum of increase in

developed land. One of its forms is expansion. Whereas urban sprawl is irregular expansion of city/town having some special characteristics (a

negative connotation).

METHODOLOGY

URBAN SPRAWL

OBJECTIVES

ANALYZE THE LAND

USE/LAND COVER

CHANGE OVER THE

LAST THREE.

STUDY THE URBAN

GROWTH PATTERN

USING SHANNON'S

ENTROPY INDEX.

PREDICT THE

URBAN GROWTH

USING THE

MODIFIED-CA

TECHNIQUE.

FORECAST THE

CHANGE IN THE

PROPERTY PRICE

IN THE YEAR 2025

To analyze the land use/land cover change over the last three

decades from multi-temporal satellite images.

To study the urban growth pattern using Shannon's entropy

index.

LITERATURE REVIEW

JOURNALS AND RESEARCHES

DATA COLLECTION

SPATIAL DATA

Accuracy

assessment

LAND USE LAND

COVER USING SUPPORT

VECTOR MACHINE

1990 2000

2010 2020

CHANGE

DETECTION

CLASSIFICATION

USING NEURAL

NETWORK

SPATIAL

ORIENTATION OF

URBAN GROWTH IN

SPIDER CHART

SHANON'S ENTROPY

MODEL SELECTION

DRIVING FACTORS

RESTRICTING

FACTORS

DELIVERABLES

LINEAR

REGRESSION

MODEL

SPATIAL

AUTOCORRELATION

Source: HCP Design, Planning and Management Pvt. Ltd

Study of urban expansion and its influence on

property price trends in Ahmedabad city and

surrounding regions

To predict the urban growth for the year 2030 using the

modified-CA technique.

To forecast the change in the property price in the year 2025 for

AMC area.

AMC database

DEM

Multispectral data

Population density

NON-SPATIAL DATA

Property Price

Change matrix

URBAN SPRAWL

PREDICTION FOR 2030

PROPERTY PRICE

FORECASTING FOR 2025

Study area

AHMEDABAD CITY PROFILE

The history of Ahmedabad can be divided into three distinct

periods, 1st The establishment during the Sultanate rule and

the pre-colonial period; 2nd, the British rule; and 3rd, the

post-colonial period.

The GHSL has been produced by the EC JRC as open and free data. These data contain a

multitemporal information layer on built-up presence as derived from Landsat image

collections (GLS1975, GLS1990, GLS2000, and ad-hoc Land sat 8 collection 2015

F o r a n a l y s i s a n d

prediction of growth in

the city and the outskirt

regions in our study

area, we took proximity

of 2.5 km from the

AMC limit and in doing

so we also avoided

influence in our study

area's growth from

n e i g h b o r i n g c i t y -

Gandhinagar. later we

created the square grid

extent from our 2.5km

buffer by minimum

bound geometry tool.

2.5 km

Latitude 23.033

Longitude 72.585

Figure 1 Pre-colonial development

The formation of city from the citadel is shown in 1st figure,

to the major industrial, commercial and residential growth in

post-colonial period is illustrated in 2nd fig below.

Study area consists of

934.32 sqkm of land

Source: Desai sowmya (2005), Urban spatial structures & land management mechanisms unpublished m. tech. planning dissertation, cept university ahmedabad

Figure 2 Development during colonial and post-colonial period

Ahmedabad is the largest city and former capital of the Indian state of Gujarat.

OBJECTIVE-1

TO ANALYZE THE LAND USE/LAND COVER CHANGE OVER THE LAST THREE DECADES FROM

MULTI-TEMPORAL SATELLITE IMAGES.

LULC 1990 LULC 2000 LULC 2010 LULC 2020

These maps are showing classification from year 1990

to 2020.

The accuracy assessment has been done with stratified

random accuracy point generation followed by

preparing confusion matrix.

LAND USE/LAND COVER CHANGE OVER THE LAST THREE DECADES

We have prepared LULC change map from 1990 to 2020 to analyze the

transition of each classes to other classes throughout these years.

Land Use Land Cover (SVM)

600.00

500.00

Area in sq.km

400.00

300.00

200.00

100.00

0.00

Built-up Agriculture / Plantation Water body Fallow land

1990 119.76 511.53 6.45 296.57

2000 199.73 354.1428 4.3875 376.0668

2010 241.056 294.5529 9.2556 389.4579

2020 310.4172 195.8418 12.9375 415.1259

Class Name

1990 2000 2010 2020

Accuracy we got for year 1990 LULC was 78.48%

followed by 2000 with 88.10% , 2010 with 84.60% and

year 2020 with 78.21% accuracy.

For built-up, the highlighted boxes are showing the

major transitions coming from other classes like

agriculture and fallow land.

The chart shows the statistics of change maps, where the areas of transitions

from 16 different classes can be known.

The change matrix has been prepared for the year 1990- 2020 to see the major transitions coming from

various classes throughout these years.

OBJECTIVE-2

TO STUDY THE URBAN GROWTH PATTERN USING SHANNON'S ENTROPY INDEX.

Classification using Neural Network - A Machine learning approach

SHANON'S ENTROPY

Shannon’s entropy describe the degree of dispersion or spatial concentration of a specific variable in a particular area.

Labelled data to train the model, this is a Supervised ML approach.

• Google’s Tensorflow library in Python to build a Neural Network (NN)

• pyrsgis — to read and write GeoTIFF

• scikit-learn — for data pre-processing and accuracy checks

• numpy — for basic array operations

B u i l d t h e m o d e l

using keras. Sequential

model to add the layers

one after the other

1990

(Bhatta(2009), Li (2004))

Shannon’s entropy (Hn) is used to measure the degree of spatial concentration or dispersion of geophysical parameter among the n zones.

The value of entropy varies from zero

to log(n). The value of zero indicates

that the distribution is very compact,

and the value near to log(n) is highly

disperse in nature. High value of

entropy indicates occurrence of

sprawl

2020

Source: Remote Sensing and Spatial Information Sciences, Volume XLII-3/W11, 2020 PECORA 21/ISRSE 38 Joint Meeting, 6–11 October 2019, Baltimore, Maryland, USA

TO PREDICT THE URBAN GROWTH FOR THE YEAR 2030 USING THE MODIFIED-CA TECHNIQUE.

LULC 2020 SIMULATION 2020 PREDICTED 2030

OBJECTIVE-3

CA model was formulated to consider all the factors which contribute to urban

growth in Study area.

The model depends primarily on the current state of the test pixel, the current state of immediate

neighboring pixels and the set of transition rules

• The simulation of year 2020

was done by considering

growth of 2000 and 2010 as a

initial years.

• The model accuracy was

85.05% with the total built-up

area of 293.73 SqKm.

Centre business district Proximity to Major Road

Amenities Quality of life Restricted zones

Assign –ve value if less than rule is required. If for a parameter, we want growth to take

place value are more than something in that case +ve value and for some parameter we

want less than something in that case value put –ve value

H e r e , H i g h

value attracts

t h e g r o w t h

t o w a r d s

features

*sqkm

• After the acceptable accuracy

from simulation result, the

prediction for the year 2030

was carried out.

• The model accuracy was

84.07% with the total built-up

area of 372..74 SqKm.

• Highly developing areas of

present time like Vaishnovdevi

and Bodakdev has shown

noticeable growth.

FORECASTING THE CHANGE IN THE PROPERTY PRICE FOR THE YEAR 2025

FORECASTED % INCREASE IN PRICE FOR 2025

OBJECTIVE-4

It is challenging to analyze or predict the property prices at

ward level as it mainly differs at micro levels, But it can be

possible to find correlation between major factors that may

affect the property prices.

Timeline Considered

• 5 years

Parameters

• Urban Growth

• Property price

.

Data Used

Very Low

• For Urban growth: Urban/built-up in 2015, 2020 and

predicted 2025 Built-up

• For Property Price: Apartment Prices in year 2015

and 2020 from the sources like: 99Acre.com,

MagicBricks.com and Makaan.com

%increase in built-up from year 2015-2020

Here, Very low value indicate below 1.72

percentage increase, Low value indicates

1.72% to 3.78%, Moderate value indicates

3.78% to 11.40%, High values indicate

11.40% to 33.24% and Rapid value indicates

more than 84.66% increase in built-up.

%increase in Price from year 2015-2020

• Almost every ward has increase in

property prices.

• Here, very low value indicate below 12

percentage increase, and Rapid value

indicates more than 118% increase in

Price.

LULC 2015 Predicted 2025

• The map shows LULC for 2015 and predicted built up for the year 2025.

• The model accuracy of 73.23 % was achieved with the built-up area of

346.18 Sqkm.

*sq km

An attempt to model the relationship between two variables by fitting a linear equation to

observed data. Where One variable is an Independent Variable, and the other is a dependent

variable.

Variables used

Independent Variable: percentage change in Built-up from 2015 to 2020 and 2020 to predicted

2025

Dependent Variable: percentage change in property prices from 2015 to 2020

Linear regression model

Scatterplot

Used to determine the

strength of the relationship

between two variables. If

there appears to be no

association between the

independent and dependent

variables, then fitting a

linear regression model to

the data probably will not

provide a useful model.

The equation for the linear

regression line is y = bx + a

From the scatterplot we got the r square value of 0.0284. which infer that there is positive

association which is not that strong among the variable.

• Highly developing

areas of present time

like ‘Chandlodia’ ,

‘Gota’ , ‘Thaltej’,

and ‘Bodakdev’ are

showing higher

increase in property

prices.

• We have made an attempt to forecast the price trend for the year 2025 with

statistical model of linear regression, having % Built-up growth as Independent

and % Increase in price as Independent Variables.

• We got the forecasted increase in the property prices by 25% to 26% for the year

2025 .

SPATIAL AUTOCORRELATION FOR BUILT-UP GROWTH AND PROPERTY PRICE

%increase in built-up from year 2015-2020

• We performed spatial autocorrelation

through global Moran's index for builtup

growth from year 2015 to 2020 to

understand correlation of Built-up over

the space.

• For percentage increase in built-up we

got,

‣ P-value = 0.25

‣ Z-score(critical value) = (-1.14)

‣ Moran’s index = (-0.14)

• Clearly indicates the spatial

autocorrelation for built-up growth is

random over the space and there is no

correlation among neighboring wards

for built-up increase.

%increase in Price from year 2015-2020

• We performed spatial autocorrelation

through global Moran's index for property

price increase from year 2015 to 2020 to

understand correlation of property price

over the space.

• For percentage increase in property price

we got,

‣ P-value = 0.079

‣ Z-score(critical value) = 1.75

‣ Moran’s index = 0.18

• Indicates the spatial autocorrelation of

property price growth has clustered

pattern over the wards with significant

correlation over the space.

CONCLUSION

CA model allows experiment to conducted on simulated system rather than the real thing. It allows the alternative scenario to be

evaluated.

CA Model had the accuracy of 85.05%. These accuracy based on different parameters has taken into consideration and pixel

transition based on input parameters.

We can consider Urban growth while forecasting the property price trends, but one might look for more numbers of independent

variables to be included in the study so that it can explain the variation in the property price more significantly.

From spatial autocorrelation we are getting spatial pattern with statistical significance like Built-up growth have random pattern

over space while property price had Clustered pattern depicting correlation over the space.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!