Household and Floor Space

Final project | Household forecasting

Authors

Xinyi Liu

Benjamin Starostka

Show the code 
import pandas as pd
import cufflinks as cf
import plotly
import plotly.express as px
import plotly.graph_objects as go
%matplotlib inline
import math
from matplotlib import pyplot as plt
from sklearn import model_selection, tree, metrics, ensemble
from sklearn.preprocessing import MinMaxScaler
from bokeh import models, plotting
from bokeh.palettes import Spectral8, Spectral11
import seaborn as sns
import pycountry as pc
from bokeh.models import HoverTool
from bokeh.plotting import figure, show
from bokeh.models import  ColumnDataSource, Legend, LinearAxis, FactorRange
from bokeh.io import output_notebook, show, curdoc
from bokeh.palettes import Spectral6, Category20b
from bokeh.transform import factor_cmap
from bokeh.models import BoxAnnotation
from bokeh.models import Title
from bokeh.layouts import layout
from bokeh.models import Tabs, TabPanel, Range1d

output_notebook()
import bokeh.models as bkm
from bokeh.resources import INLINE
import bokeh.io
from bokeh import *
bokeh.io.output_notebook(INLINE)
import regex as re

from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot 
init_notebook_mode(connected=True)
cf.go_offline()
Loading BokehJS ...
Loading BokehJS ...

1 Growth Overview

With the sharp increase in the number of population, households are also expected to grow in a similar pattern in order to accommodate the population. Based on the rapid growth of population, we are able to observe that in next 20 years, the number of households and floor space will increase accordingly.

Show the code 
df_population = pd.read_csv('../data/city-of-melbourne-population-forecasts-by-small-area-2020-2040.csv')
df_household = pd.read_csv('../data/city-of-melbourne-dwellings-and-household-forecasts-by-small-area-2020-2040.csv')
df_floorspace = pd.read_csv('../data/city-of-melbourne-floor-space-forecasts-by-small-area-2020-2040.csv')

df_age = df_population.drop(df_population[df_population['Age'].isin(['Total population', 'Average age'])].index)
df_age = df_age.drop(df_age[df_age['Geography'].isin(['City of Melbourne'])].index)

df_household = df_household.drop(df_household[df_household['Geography'].isin(['City of Melbourne'])].index)
df_household = df_household.drop(df_household[df_household['Households'].isin(['Not applicable', 'Total households'])].index)

df_floorspace = df_floorspace.drop(df_floorspace[df_floorspace['Geography'].isin(['City of Melbourne'])].index)
df_floorspace = df_floorspace.drop(df_floorspace[df_floorspace['Industry Space Use'].isin(['Total residential floorspace'])].index)
df_floorspace = df_floorspace.drop(df_floorspace[df_floorspace['Industry Space Use'].isin(['Total employment floorspace'])].index)

households_by_year = df_household.groupby(['Year']).sum()
households_by_year = households_by_year['Value'].to_frame().reset_index()
floor_by_year = df_floorspace.groupby(['Year']).sum()
floor_by_year = floor_by_year['Value'].to_frame().reset_index()
df_age = df_age.groupby(['Year']).sum()
df_age = df_age['Value'].to_frame().reset_index()

forecast_overview = pd.merge(households_by_year, floor_by_year, on='Year')
forecast_overview = pd.merge(forecast_overview, df_age, on='Year')

forecast_overview.rename(columns={'Value_x': 'Households', 'Value_y': 'Floor Space', 'Value': 'Population'}, inplace=True)

columns_to_normalize = ['Households', 'Floor Space', 'Population']

# Initialize the scaler
scaler = MinMaxScaler()
normalized_data = scaler.fit_transform(forecast_overview[columns_to_normalize])

df_normalized = pd.DataFrame(normalized_data, columns=columns_to_normalize)
df_normalized = pd.concat([forecast_overview['Year'], df_normalized], axis=1)

# Viz the normalized data
df_long = pd.melt(df_normalized, id_vars='Year', var_name='Category', value_name='Value')

# Create the bar plot using Plotly Express
fig = px.bar(df_long, x='Year', y='Value', color='Category',
             title='Population, Dwelling and Floor Space Forecast Overview (Normalized)',
             labels={'Year': 'Year', 'Value': 'Value'}, barmode='group')

fig.update_layout(width=600, height=500, title_font_size=13)

# Show the plot
fig.show()

The growth of households and floor space reflects the fact that Melbourne is a pleasant city for people to live in. The municipality of Melbourne has dedicated efforts to enhance the residents’ quality of life. By investing in infrastructure and public amenities, the city aims to create a more livable environment. How about different small areas in Melbourne? Are they all growing at the same pace?

Show the code 
forecast_overview_viz = pd.merge(forecast_overview, df_normalized, on='Year')
forecast_overview_viz.rename(columns={'Households_x': 'Households', 'Floor Space_x': 'Floor_Space', 'Population_x': 'Population', 'Households_y': 'Households (Normalized)', 'Floor Space_y': 'Floor_Space (Normalized)', 'Population_y': 'Population (Normalized)'}, inplace=True)

source = models.ColumnDataSource(forecast_overview_viz)

p1 = plotting.figure(title='Melbourne Forecast Overview (Normalized)', x_axis_label='Year', y_axis_label='Population', width=700, height=350)

plot1 =p1.line(x='Year', y='Population (Normalized)', source=source, legend_label='Population', line_width=3, color=Spectral11[1])
plot2 =p1.line(x='Year', y='Households (Normalized)', source=source, legend_label='Households', line_width=3, color=Spectral11[3])
plot3 =p1.line(x='Year', y='Floor_Space (Normalized)', source=source, legend_label='Floor_Space', line_width=3, color=Spectral11[5])

p1.add_tools(models.HoverTool(renderers=[plot1], tooltips=[('Year', '@Year'), ('Population', '@Population')]))
p1.add_tools(models.HoverTool(renderers=[plot2], tooltips=[('Year', '@Year'), ('Households', '@Households')]))
p1.add_tools(models.HoverTool(renderers=[plot3], tooltips=[('Year', '@Year'), ('Floor Space', '@Floor_Space sqm')]))

p1.legend.location = 'top_left'
p1.legend.background_fill_alpha = 0.25
p1.legend.background_fill_color = 'darkgray'
p1.legend.border_line_alpha = 0
p1.title.align = 'center'


df_household.rename(columns={'Value': 'Number of Households'}, inplace=True)
df_household_area = df_household.groupby(['Year', 'Geography']).sum()
# df_household_area.drop(columns=['Households', 'Category'], inplace=True)
df_household_area.reset_index(inplace=True)
df_household_area

df_household_area = df_household_area.pivot(index='Year', columns='Geography', values='Number of Households')

source = models.ColumnDataSource(df_household_area)

areas = df_household['Geography'].unique().tolist()
years = df_household['Year'].unique().tolist()
years = years.sort()

p2 = figure(title='Melbourne Households Forecast by Small Area', x_axis_label='Year', y_axis_label='Number of Households', width=700, height=350)

bar = {}
for indx, i in enumerate(areas):
    bar[i] = p2.vbar(x="Year", top=i, width=0.3, 
                    source=source, muted_alpha=0.2, muted=False, color=Category20b[14][indx])

legend = Legend(items=[(x, [bar[x]]) for x in areas], location=(0, -30))
p2.add_layout(legend, 'right')

p2.legend.background_fill_alpha = 0.25
p2.legend.border_line_alpha = 0
p2.title.align = 'center'
p2.legend.click_policy = 'hide'

df_floorspace.rename(columns={'Value': 'Area of Floor Space'}, inplace=True)
df_floorspace_area = df_floorspace.groupby(['Year', 'Geography']).sum()
# df_floorspace_area.drop(columns=['Industry Space Use', 'Category'], inplace=True)
df_floorspace_area.reset_index(inplace=True)
df_floorspace_area = df_floorspace_area.pivot(index='Year', columns='Geography', values='Area of Floor Space')

source = models.ColumnDataSource(df_floorspace_area)

areas = df_floorspace['Geography'].unique().tolist()
years = df_floorspace['Year'].unique().tolist()
years = years.sort()

p3 = figure(title='Melbourne Floor Space Forecast by Small Area', x_axis_label='Year', y_axis_label='Area of Floor Space', width=700, height=350)

bar = {}
for indx, i in enumerate(areas):
    bar[i] = p3.vbar(x="Year", top=i, width=0.3, 
                    source=source, muted_alpha=0.2, muted=False, color=Category20b[14][indx])

legend = Legend(items=[(x, [bar[x]]) for x in areas], location=(0, -30))
p3.add_layout(legend, 'right')

p3.legend.background_fill_alpha = 0.25
p3.legend.border_line_alpha = 0
p3.title.align = 'center'
p3.legend.click_policy = 'hide'


tab1 = TabPanel(child=p1,title="Forecast Overview")
tab2 = TabPanel(child=p2,title="Households by Small Area")
tab3 = TabPanel(child=p3,title="Floor Space by Small Area")

tabs = Tabs(tabs=[tab1, tab2, tab3])

curdoc().add_root(tabs)

show(tabs)

By 2041, the City of Melbourne is forecast to reach 136,000 households, accommodating an additional 50,000 households between 2021 and 2041. Floor space demand is expected to grow by over 6.3 million square metres between 2021 and 2041, of which 2.7 million square metres is residential.

Melbourne (CBD) dominates the floor space within the Melbourne municipality, showcasing the most significant projected growth over the next 20 years. North Melbourne, Southbank, Docklands and Carlton are the other suburbs that are expected to have a significant increase by the amount of floor space. The rest of areas remains relatively stable.

Despite of Melbourne CBD, Carlton, North Melbourne and Southbank are predicted to have a notable increase in the number of households. The prediction may be due to the rapid development and resource allocation in these areas, rendering them more appealing as residential destinations. The reason of Mebelbourne dominating the floor space is that it is the central business district of Melbourne, which is the main commercial and administrative centre of the city, with the most convenient transportation and the most prosperous business.

2 Takeaways

By looking at the households and floor space forecast data in Merbourne, we are able to conclue that: - Over the next twenty years, Melbourne CBD is expected to retain its status as the city center with the highest population density and the largest number of households. - The expansion of households and floor space is failing to keep pace with the growth of the population, therefore more space in the suburb will be developed to accommadate more and more people in Melbourne.

Reuse

CC