Brussels Mat. Step 2.a Construction Aggregates¶

In [1]:

import pandas as pd
import numpy as np
from matplotlib import pyplot as plt

In [62]:

constructions = pd.read_excel('IBSA/11.2_amenagement_territoire_parc_batiments_20170626.xlsx',
                    sheetname='11.2.5.1', skiprows = 1, #header = 0,
                    index_col = 0, skip_footer = 10)
constructions.loc[:, 'region'] = constructions.index
constructions = constructions.set_index('Année')

In [63]:

constructions.head()

Out[63]:

	Bâtiments résidentiels	Unnamed: 3	Unnamed: 4	Unnamed: 5	Unnamed: 6	Unnamed: 7	Unnamed: 8	Bâtiments non résidentiels	Unnamed: 10	Unnamed: 11	Unnamed: 12	Unnamed: 13	region
Année
NaN	Nouvelles constructions	NaN	NaN	NaN	NaN	Rénovation	NaN	Nouvelles constructions	NaN	NaN	Rénovation	NaN	NaN
NaN	Nombre de bâtiments	Nombre de logements	Nombre d'appartements	Nombre de bâtiments avec un seul logement	Superficie habitable (m²)	Nombre de bâtiments	Nombre de logements	Nombre de bâtiments	Nombre de logements	Volume (m³)	Nombre de bâtiments	Nombre de logements	Colonne3
1996.0	234	1375	1231	144	167147	960	2264	77	36	1670453	167	117.938	Région de Bruxelles-Capitale
1996.0	3875	4693	1003	3690	723710	2102	2389	489	64	2378497	367	43.015	Brabant flamand
1996.0	1324	2546	1319	1227	308394	949	1460	177	24	724580	143	27.3	Brabant wallon

In [56]:

def get_new_cols(iterable, keyword):
    if not isinstance(keyword, str):
        keyword = 'NaN'
    new_cols_0 = list()
    old_cols = str()
    for i in iterable:
        if not isinstance(i, str):
            i = 'NaN'
        if keyword in i or keyword == i:
            new_cols_0.append(old_cols)
        else:
            new_cols_0.append(i)
            old_cols = i
    return(new_cols_0)

In [57]:

new_cols_0 = get_new_cols(constructions.columns, 'Unnamed')
new_cols_1 = get_new_cols(constructions.iloc[0,:], np.nan)
new_cols_2 = get_new_cols(constructions.iloc[1,:], np.nan)

In [59]:

constructions.columns = [new_cols_0, new_cols_1, new_cols_2]
constructions = constructions.iloc[2:, :]
inx = constructions.region == 'Région de Bruxelles-Capitale'
constructions = constructions.loc[inx.iloc[:,0]]
constructions = constructions.iloc[:,:-1]

In [6]:

constructions = constructions.T.sort_index()

In [7]:

sqm = constructions.loc[(
    slice(None),
    #'Bâtiments résidentiels',
    slice(None),
    #slice('Nouvelles constructions'),
    [
        #'Superficie habitable (m²)',
        #'Nombre de logements',
        'Nombre de bâtiments'
    ],
), :]

In [8]:

sqm.index = sqm.index.droplevel([-1])
sqm_per = pd.concat([
    pd.DataFrame(sqm.iloc[0].div(sqm.iloc[1] + sqm.iloc[0]).mul(100), columns=['Non Residential']),
    pd.DataFrame(sqm.iloc[2].div(sqm.iloc[3] + sqm.iloc[2]).mul(100), columns=['Residential'])],
    axis=1)

In [9]:

g = sqm_per.plot(figsize=(10, 5))
g.set_ylabel(r'Share of new constructions $[\%]$');
g.set_title("Share of new constructed buildings of total constructed buildings");
for i in [0, 1]:
    year_max = sqm_per.index[sqm_per.iloc[:, i] == sqm_per.iloc[:, i].max()][0]
    year_min = sqm_per.index[sqm_per.iloc[:, i] == sqm_per.iloc[:, i].min()][0]
    g.text(year_max, sqm_per.iloc[:, i].max(),
           "{1:0.2f} $\%$\n{0:0.0f} (max)".format(year_max, sqm_per.iloc[:, i].max()),
           style='italic', horizontalalignment='right',
           bbox={'facecolor':'red', 'alpha':0.8, 'pad':2})
    g.text(year_min, sqm_per.iloc[:, i].min(),
           "{1:0.2f} $\%$\n{0:0.0f} (min)".format(year_min, sqm_per.iloc[:, i].min()),
           style='italic', horizontalalignment='right',
           bbox={'facecolor':'red', 'alpha':0.8, 'pad':2})
g.grid()
plt.savefig('FIGURES/btyp_share_new.png', dpi=300)

../../_images/example_be2_Ba_Construction_Aggregates_10_0.png

In [10]:

sqm_tot = pd.concat([
    pd.DataFrame(sqm.iloc[1] + sqm.iloc[0], columns=['Non Residential']),
    pd.DataFrame(sqm.iloc[3] + sqm.iloc[2], columns=['Residential'])],
    axis=1)

In [11]:

sqm_tot.div(sqm_tot.iloc[0]).iloc[-1]

Out[11]:

Non Residential    0.647541
Residential         1.74372
Name: 2013.0, dtype: object

In [12]:

g = sqm_tot.div(sqm_tot.iloc[0]).plot(figsize=(10, 5))
g.set_ylabel(r'Index buildings constructed');
g.set_title("Index of number of constructed buildings (1996=1)");
for i in [0, 1]:
    g.text(2013, sqm_tot.div(sqm_tot.iloc[0]).iloc[-1, i],
           "{0:0.0f} buildings (2013 = {1:0.2f})".format(sqm_tot.iloc[-1, i],
                                               sqm_tot.div(sqm_tot.iloc[0]).iloc[-1, i]),
           style='italic', horizontalalignment='right',
           bbox={'facecolor':'red', 'alpha':0.8, 'pad':2})
g.grid()
plt.savefig('FIGURES/btyp_total_buildings.png', dpi=300)

../../_images/example_be2_Ba_Construction_Aggregates_13_0.png

In [13]:

sqm_du = constructions.loc[(
    'Bâtiments résidentiels',
    slice('Nouvelles constructions'),
    [
        'Superficie habitable (m²)',
        'Nombre de logements',
    ],
), :]
sqm_du.index = sqm_du.index.droplevel([0,1])

In [14]:

sqm_du = sqm_du.iloc[1].div(sqm_du.iloc[0])
#sqm_du.name = 'sqm per DU'

In [49]:

MU_du = 74.92
MU_sd = 7.01
x = [i for i in range(2001, 2014)]
y = [MU_du + i/4 for i in range(13)]
y1 = [i+MU_sd for i in y]
y2 = [i-MU_sd for i in y]

In [53]:

g = sqm_du.plot(figsize=(10, 5), label='new DU')
g.set_ylabel('$m^2$ per dwelling unit')
g.set_xlabel('year')
g.set_title("Average dwelling unit size for new constructions (residential)")
for e, i in enumerate(sqm_du.index):
    if e in [0, 3, 6, 12, 16, 17]:
        g.text(i, sqm_du.loc[i],
               "{0:0.0f} = {1:0.2f} $m^2$".format(i, sqm_du.loc[i]),
               style='italic', horizontalalignment='right',
               bbox={'facecolor':'red', 'alpha':0.8, 'pad':2})
g.grid()
g.plot(x, y, "--", label='existing DU $\mu$')
g.fill_between(x, y1=y1, y2=y2,
               label='existing DU $\sigma$', color='green', alpha=0.3)
g.legend()
plt.savefig('FIGURES/btyp_du_size.png', dpi=300)

../../_images/example_be2_Ba_Construction_Aggregates_17_0.png