2 - Where to spend leisure time in NYC?

This project aims to explore the spatial distribution of some leisure spaces in New York City, including indoor spaces like museums, art galleries, and theatres, and outddoor spaces like parks or open streets. The project first aims to idenfity some patterns of spatial distribution of such spaces in each borough and neighborhood, and in relation to population and median household income. Then some interactive visualization is created for both local NYC residents and tourists to access to information about some leisure spaces more easily.

Imports

Code

import altair as alt
import geopandas as gpd
import hvplot.pandas
import numpy as np
import pandas as pd
import seaborn as sns
from matplotlib import pyplot as plt
import datetime
import math

%matplotlib inline

NYC Base Maps

Tracts, Neighborhood, and Borough

nyc_tracts = pd.read_csv(“data/2022 Census Tracts.csv”) nyc_tracts[‘geometry’] = gpd.GeoSeries.from_wkt(nyc_tracts[‘the_geom’]) geo_tracts = gpd.GeoDataFrame(nyc_tracts, geometry=‘geometry’) geo_tracts = geo_tracts.set_crs(epsg=4326) tracts = geo_tracts[[‘BoroName’, ‘CT2020’, ‘BoroCT2020’, ‘NTAName’, ‘Shape_Area’,‘geometry’,‘GEOID’]].copy() tracts_clean = tracts[[‘BoroName’, ‘CT2020’, ‘NTAName’, ‘geometry’,‘GEOID’]] Boro_NTA = tracts_clean[[‘BoroName’, ‘NTAName’]].drop_duplicates(subset = “NTAName”)

Dataset Selection & Setup

After glancing through NYC Open Data portal, I have selected art galleries, museums, libraries, theatres, and parks as common leisure spaces. In addition, I explored data of open streets, a program thatt had been implemented in many global cities including New York City to transform roads into public spaces for cultural and all kinds of events on particular days (mostly on weekends).

This step is to bring in all datasets, clean them up, and aggregate different types of leisure spaces. The aggregated dataframe is then spacially joined with tracts, neighborhood, and boroughs, ready for further investigation on their geospatial distributions.

Similar data wrangling is performed on parks. Since parks come in polygon instead of points, which may result in problems with spatial joins, in the case when one park falls in two or more tracts or neighborhood. Hence, the geometry of parks’ centroids is adopted to replace the polygon geometry for further analysis.

The Open Street data comes with more detailed information on the approved time for each street. The days of week (e.g. Monday) and time of the day (e.g. 7:30) that it opens and closes are rearranged by melting and pivoting.

Additionally, 2020 Census Data on population and median household income are brought in for analysis.

Indoor Leisure Spaces: Art Galleries, Museums, Libraries, and Theatres

Code

# establish geodataframe
art_galleries = gpd.read_file("data/art galleries.geojson")
geo_art_galleries = gpd.GeoDataFrame(art_galleries, geometry='geometry')
geo_art_galleries = art_galleries.set_crs(epsg=4326)

# add type
art_galleries_clean = geo_art_galleries[['name','zip','address1','geometry']]
art_galleries_clean.loc[:,"Type"]= "Art Galleries"
art_galleries_clean.rename(
    columns={"address1": "Address", "name": "Name", "zip": "Zip"},
    inplace=True,)

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1429971093.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  art_galleries_clean.rename(

Code

# Museums 

# establish geodataframe
museums = gpd.read_file("data/museums.geojson")
geo_museums= gpd.GeoDataFrame(museums, geometry='geometry')
geo_museums = geo_museums.set_crs(epsg=4326)
geo_museums

# add type
museums_clean = geo_museums[['name','zip','adress1','geometry']]
museums_clean.loc[:,"Type"]= "Museums"
museums_clean.rename(
    columns={"adress1": "Address", "name": "Name", "zip": "Zip"},
    inplace=True,)
museums_clean

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/3071166290.py:12: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  museums_clean.rename(

	Name	Zip	Address	geometry	Type
0	Alexander Hamilton U.S. Custom House	10004.0	1 Bowling Grn	POINT (-74.01376 40.70382)	Museums
1	Alice Austen House Museum	10305.0	2 Hylan Blvd	POINT (-74.06303 40.61512)	Museums
2	American Academy of Arts and Letters	10032.0	633 W. 155th St.	POINT (-73.94730 40.83385)	Museums
3	American Folk Art Museum	10019.0	45 West 53rd Street	POINT (-73.97810 40.76162)	Museums
4	American Immigration History Center	0.0	Ellis Island	POINT (-74.03968 40.69906)	Museums
...	...	...	...	...	...
125	American Sephardi Federation / Sephardic House	10011.0	15 W. 16th St.	POINT (-73.99389 40.73808)	Museums
126	YIVO Institute for Jewish Research	10011.0	15 W. 16th St.	POINT (-73.99379 40.73796)	Museums
127	American Jewish Historical Society	10011.0	15 W. 16th St.	POINT (-73.99393 40.73802)	Museums
128	Yeshiva University Museum	10011.0	15 W. 16th St.	POINT (-73.99382 40.73805)	Museums
129	Center For Jewish History	10011.0	15 W. 16th St.	POINT (-73.99387 40.73799)	Museums

130 rows × 5 columns

Code

# Libraries

# establish geodataframe
library = gpd.read_file("data/libraries.geojson")
geo_library = gpd.GeoDataFrame(library, geometry='geometry')
geo_library= geo_library.set_crs(epsg=4326)
geo_library

# add type
library_clean = geo_library[['name','zip','streetname','geometry']]
library_clean.loc[:,"Type"]= "Libraries"
library_clean.rename(
    columns={"streetname": "Address", "name": "Name", "zip": "Zip"},
    inplace=True,)
library_clean

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/3524914390.py:12: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  library_clean.rename(

	Name	Zip	Address	geometry	Type
0	115th Street	10026	West 115th Street	POINT (-73.95353 40.80298)	Libraries
1	125th Street	10035	East 125th Street	POINT (-73.93485 40.80302)	Libraries
2	53rd Street	10019	West 53rd Street	POINT (-73.97736 40.76081)	Libraries
3	58th Street	10022	East 58th Street	POINT (-73.96938 40.76219)	Libraries
4	67th Street	10065	East 67th Street	POINT (-73.95955 40.76492)	Libraries
...	...	...	...	...	...
211	Sunnyside	11104	Greenpoint Avenue	POINT (-73.92167 40.74085)	Libraries
212	Whitestone	11357	14 Road	POINT (-73.81070 40.78854)	Libraries
213	Windsor Park	11364	Bell Boulevard	POINT (-73.75562 40.73450)	Libraries
214	Woodhaven	11421	Forest Parkway	POINT (-73.86146 40.69453)	Libraries
215	Woodside	11377	Skillman Avenue	POINT (-73.90979 40.74534)	Libraries

216 rows × 5 columns

Code

# Theaters
# establish geodataframe
theaters = gpd.read_file("data/Theaters.geojson")
geo_theaters = gpd.GeoDataFrame(theaters, geometry='geometry')
geo_theaters= geo_theaters.set_crs(epsg=4326)
geo_theaters


# add type
theaters_clean = geo_theaters[['name','zip','address1','geometry']]
theaters_clean.loc[:,"Type"]= "Theatres"
theaters_clean.rename(
    columns={"address1": "Address", "name": "Name", "zip": "Zip"},
    inplace=True,)
theaters_clean

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1518346627.py:12: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  theaters_clean.rename(

	Name	Zip	Address	geometry	Type
0	45th Street Theater	10036.0	354 West 45th Street	POINT (-73.99062 40.75985)	Theatres
1	47th Street Theater	10036.0	304 West 47th Street	POINT (-73.98811 40.76047)	Theatres
2	59E59	10022.0	59 East 59th Street	POINT (-73.97038 40.76340)	Theatres
3	Acorn Theater	10036.0	410 West 42nd Street	POINT (-73.99332 40.75854)	Theatres
4	Al Hirschfeld Theater	10036.0	302 W 45th Street	POINT (-73.98921 40.75926)	Theatres
...	...	...	...	...	...
112	Westside Theater	10036.0	407 W 43rd St	POINT (-73.99255 40.75953)	Theatres
113	Wings Theatre	10014.0	154 Christopher St	POINT (-74.00889 40.73240)	Theatres
114	Winter Garden Theatre	10019.0	1634 Broadway	POINT (-73.98348 40.76152)	Theatres
115	York Theatre	10022.0	619 Lexington Ave	POINT (-73.96998 40.75836)	Theatres
116	Delacorte Theater	0.0	Central Park - Mid-Park at 80th Street	POINT (-73.96882 40.78018)	Theatres

117 rows × 5 columns

Aggregate all types of indoor leisure space

Code

frames = [art_galleries_clean, museums_clean, library_clean, theaters_clean]

total_indoor = pd.concat(frames)

total_indoor

	Name	Zip	Address	geometry	Type
0	O'reilly William & Co Ltd	10021.0	52 E 76th St	POINT (-73.96273 40.77380)	Art Galleries
1	Organization of Independent Artists - Gallery 402	10013.0	19 Hudson St.	POINT (-74.00939 40.71647)	Art Galleries
2	Owen Gallery	10021.0	19 E 75th St	POINT (-73.96435 40.77400)	Art Galleries
3	P P O W Gallerie	10001.0	511 W 25th St	POINT (-74.00389 40.74959)	Art Galleries
4	P P O W Inc	10013.0	476 Broome St	POINT (-74.00176 40.72291)	Art Galleries
...	...	...	...	...	...
112	Westside Theater	10036.0	407 W 43rd St	POINT (-73.99255 40.75953)	Theatres
113	Wings Theatre	10014.0	154 Christopher St	POINT (-74.00889 40.73240)	Theatres
114	Winter Garden Theatre	10019.0	1634 Broadway	POINT (-73.98348 40.76152)	Theatres
115	York Theatre	10022.0	619 Lexington Ave	POINT (-73.96998 40.75836)	Theatres
116	Delacorte Theater	0.0	Central Park - Mid-Park at 80th Street	POINT (-73.96882 40.78018)	Theatres

1380 rows × 5 columns

Spatial Join

Code

geo_total_indoor = gpd.sjoin(
    total_indoor,  # The point data for 311 tickets
    tracts_clean.to_crs(total_indoor.crs),  # The neighborhoods (in the same CRS)
    predicate="within",
    how="left",
)
geo_total_indoor.head()

	Name	Zip	Address	geometry	Type	index_right	BoroName	CT2020	NTAName	GEOID
0	O'reilly William & Co Ltd	10021.0	52 E 76th St	POINT (-73.96273 40.77380)	Art Galleries	84.0	Manhattan	13000.0	Upper East Side-Carnegie Hill	3.606101e+10
1	Organization of Independent Artists - Gallery 402	10013.0	19 Hudson St.	POINT (-74.00939 40.71647)	Art Galleries	17.0	Manhattan	3900.0	Tribeca-Civic Center	3.606100e+10
2	Owen Gallery	10021.0	19 E 75th St	POINT (-73.96435 40.77400)	Art Galleries	84.0	Manhattan	13000.0	Upper East Side-Carnegie Hill	3.606101e+10
3	P P O W Gallerie	10001.0	511 W 25th St	POINT (-74.00389 40.74959)	Art Galleries	1134.0	Manhattan	9901.0	Chelsea-Hudson Yards	3.606101e+10
4	P P O W Inc	10013.0	476 Broome St	POINT (-74.00176 40.72291)	Art Galleries	1156.0	Manhattan	4900.0	SoHo-Little Italy-Hudson Square	3.606100e+10

Outdoor Leisure Space: Parks

Code

# Parks
# establish geodataframe
parks = gpd.read_file("data/Parks Properties.geojson")
geo_parks = gpd.GeoDataFrame(parks, geometry='geometry')
geo_parks= geo_parks.set_crs(epsg=4326)
geo_parks


# add type
parks_clean = geo_parks[['signname','geometry','acres']]
parks_clean.loc[:,"Type"]= "Parks"
parks_clean.rename(
    columns={"location": "Address", "signname": "Name", "zipcode": "Zip"},
    inplace=True,)

parks_clean['new_geom'] = parks_clean['geometry']

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1177565529.py:12: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  parks_clean.rename(
/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)

Code

# Since many parks are in multi-polygons, which will lead to errors when doing spatial join later, a centroid of each park is generated here for smoother spatial join.
parks_clean['geometry'] = parks_clean['geometry'].centroid

geo_parks_clean = gpd.sjoin(
    parks_clean,  # The point data for 311 tickets
    tracts_clean.to_crs(parks_clean.crs),  # The neighborhoods (in the same CRS)
    predicate="within",
    how="left",
)
geo_parks_clean

/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/3082874007.py:2: UserWarning: Geometry is in a geographic CRS. Results from 'centroid' are likely incorrect. Use 'GeoSeries.to_crs()' to re-project geometries to a projected CRS before this operation.

  parks_clean['geometry'] = parks_clean['geometry'].centroid
/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/geopandas/geodataframe.py:1538: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  super().__setitem__(key, value)

	Name	geometry	acres	Type	new_geom	index_right	BoroName	CT2020	NTAName	GEOID
0	Inwood Hill Park	POINT (-73.92544 40.87257)	196.398	Parks	MULTIPOLYGON (((-73.92093 40.86999, -73.92145 ...	2191.0	Manhattan	29700.0	Inwood Hill Park	3.606103e+10
1	Challenge Playground	POINT (-73.72796 40.75662)	2.035	Parks	MULTIPOLYGON (((-73.72738 40.75605, -73.72783 ...	936.0	Queens	152902.0	Douglaston-Little Neck	3.608115e+10
2	Sunset Cove Park	POINT (-73.82300 40.59853)	9.375	Parks	MULTIPOLYGON (((-73.82218 40.59892, -73.82221 ...	1129.0	Queens	107201.0	Breezy Point-Belle Harbor-Rockaway Park-Broad ...	3.608111e+10
3	Grand Central Parkway Extension	POINT (-73.85317 40.75316)	249.389	Parks	MULTIPOLYGON (((-73.85875 40.76741, -73.85976 ...	625.0	Queens	39902.0	North Corona	3.608104e+10
4	Idlewild Park	POINT (-73.75229 40.65043)	180.85	Parks	MULTIPOLYGON (((-73.75809 40.65427, -73.75845 ...	1013.0	Queens	66404.0	Springfield Gardens (South)-Brookville	3.608107e+10
...	...	...	...	...	...	...	...	...	...	...
2040	Maria Hernandez Park	POINT (-73.92386 40.70317)	6.873	Parks	MULTIPOLYGON (((-73.92251 40.70351, -73.92381 ...	1729.0	Brooklyn	42900.0	Bushwick (West)	3.604704e+10
2041	Crotona Parkway Malls	POINT (-73.88477 40.84405)	8.75	Parks	MULTIPOLYGON (((-73.88496 40.84470, -73.88496 ...	336.0	Bronx	36300.0	West Farms	3.600504e+10
2042	Park	POINT (-73.89807 40.84408)	0.511	Parks	MULTIPOLYGON (((-73.89759 40.84410, -73.89773 ...	1229.0	Bronx	16500.0	Claremont Village-Claremont (East)	3.600502e+10
2043	Cunningham Park	POINT (-73.76880 40.73382)	358.0	Parks	MULTIPOLYGON (((-73.77466 40.72442, -73.77439 ...	2241.0	Queens	128300.0	Cunningham Park	3.608113e+10
2044	Roberto Clemente Ballfield	POINT (-73.96767 40.70635)	1.93	Parks	MULTIPOLYGON (((-73.96761 40.70581, -73.96735 ...	2220.0	Brooklyn	54500.0	South Williamsburg	3.604705e+10

2045 rows × 10 columns

Outdoor Leisure Space: Open Streets

Code

open_streets = gpd.read_file("data/Open Streets Locations.geojson")
geo_open_streets = gpd.GeoDataFrame(open_streets, geometry='geometry', crs=2263)
geo_open_streets = open_streets.to_crs(epsg=4326)

open_streets_clean = geo_open_streets[['appronstre','apprtostre','apprdayswe','boroughname','reviewstat','shape_stle','geometry']]

Code

join_test = gpd.sjoin(
    open_streets_clean,  # The point data for 311 tickets
    tracts_clean.to_crs(open_streets_clean.crs),  # The neighborhoods (in the same CRS)
    predicate="within",
    how="left",
)
join_test.head()

	appronstre	apprtostre	apprdayswe	boroughname	reviewstat	shape_stle	geometry	index_right	BoroName	CT2020	NTAName	GEOID
0	DEISIUS STREET	STECHER STREET	mon,tue,wed,thu,fri	Staten Island	approvedFullSchools	264.932398036	MULTILINESTRING ((-74.18738 40.53028, -74.1882...	1318.0	Staten Island	17600.0	Annadale-Huguenot-Prince's Bay-Woodrow	3.608502e+10
1	SUFFOLK AVENUE	HAROLD STREET	fri,sat,sun	Staten Island	approvedFull	313.087821487	MULTILINESTRING ((-74.12784 40.60288, -74.1277...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10
2	SUFFOLK AVENUE	HAROLD STREET	fri,sat,sun	Staten Island	approvedFull	142.063500219	MULTILINESTRING ((-74.12772 40.60202, -74.1276...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10
3	VERMONT COURT	SUFFOLK AVENUE	fri,sat,sun	Staten Island	approvedFull	421.392020366	MULTILINESTRING ((-74.12620 40.60209, -74.1277...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10
4	9 STREET	ROSE AVENUE	fri,sat,sun	Staten Island	approvedFull	448.103939286	MULTILINESTRING ((-74.11481 40.57316, -74.1157...	1300.0	Staten Island	13400.0	New Dorp-Midland Beach	3.608501e+10

Code

join_test['monday'] = join_test['apprdayswe'].str.count('mon')
join_test['tuesday'] = join_test['apprdayswe'].str.count('tue')
join_test['wednesday'] = join_test['apprdayswe'].str.count('wed')
join_test['thursday'] = join_test['apprdayswe'].str.count('thu')
join_test['friday'] = join_test['apprdayswe'].str.count('fri')
join_test['saturday'] = join_test['apprdayswe'].str.count('sat')
join_test['sunday'] = join_test['apprdayswe'].str.count('sun')
join_test['dayscount'] = join_test[['monday', 'tuesday', 'wednesday', 'thursday', 'friday', 'saturday', 'sunday']].sum(axis=1)

join_test

	appronstre	apprtostre	apprdayswe	boroughname	reviewstat	shape_stle	geometry	index_right	BoroName	CT2020	NTAName	GEOID	monday	tuesday	wednesday	thursday	friday	saturday	sunday	dayscount
0	DEISIUS STREET	STECHER STREET	mon,tue,wed,thu,fri	Staten Island	approvedFullSchools	264.932398036	MULTILINESTRING ((-74.18738 40.53028, -74.1882...	1318.0	Staten Island	17600.0	Annadale-Huguenot-Prince's Bay-Woodrow	3.608502e+10	1	1	1	1	1	0	0	5
1	SUFFOLK AVENUE	HAROLD STREET	fri,sat,sun	Staten Island	approvedFull	313.087821487	MULTILINESTRING ((-74.12784 40.60288, -74.1277...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10	0	0	0	0	1	1	1	3
2	SUFFOLK AVENUE	HAROLD STREET	fri,sat,sun	Staten Island	approvedFull	142.063500219	MULTILINESTRING ((-74.12772 40.60202, -74.1276...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10	0	0	0	0	1	1	1	3
3	VERMONT COURT	SUFFOLK AVENUE	fri,sat,sun	Staten Island	approvedFull	421.392020366	MULTILINESTRING ((-74.12620 40.60209, -74.1277...	1323.0	Staten Island	18703.0	Todt Hill-Emerson Hill-Lighthouse Hill-Manor H...	3.608502e+10	0	0	0	0	1	1	1	3
4	9 STREET	ROSE AVENUE	fri,sat,sun	Staten Island	approvedFull	448.103939286	MULTILINESTRING ((-74.11481 40.57316, -74.1157...	1300.0	Staten Island	13400.0	New Dorp-Midland Beach	3.608501e+10	0	0	0	0	1	1	1	3
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
361	BECK STREET	AVENUE ST JOHN	wed	Bronx	approvedFull	619.314479576	MULTILINESTRING ((-73.90222 40.81449, -73.9014...	200.0	Bronx	8300.0	Longwood	3.600501e+10	0	0	1	0	0	0	0	1
362	NEWKIRK AVENUE	EAST 17 STREET	sun	Brooklyn	approvedFull	284.786787549	MULTILINESTRING ((-73.96422 40.63510, -73.9641...	1807.0	Brooklyn	52000.0	Flatbush (West)-Ditmas Park-Parkville	3.604705e+10	0	0	0	0	0	0	1	1
363	18 STREET	4 AVENUE	mon,tue,wed,thu,fri	Brooklyn	approvedFullSchools	768.660222317	MULTILINESTRING ((-73.99201 40.66323, -73.9916...	2248.0	Brooklyn	14300.0	Sunset Park (West)	3.604701e+10	1	1	1	1	1	0	0	5
364	34 AVENUE	JUNCTION BOULEVARD	mon,tue,wed,thu,fri,sat,sun	Queens	approvedLimited	272.995599601	MULTILINESTRING ((-73.89716 40.75246, -73.8970...	2168.0	Queens	29100.0	Jackson Heights	3.608103e+10	1	1	1	1	1	1	1	7
365	DECATUR STREET	SARATOGA AVENUE	sat	Brooklyn	approvedLimited	769.964464111	MULTILINESTRING ((-73.92001 40.68299, -73.9176...	1682.0	Brooklyn	37700.0	Bedford-Stuyvesant (East)	3.604704e+10	0	0	0	0	0	1	0	1

366 rows × 20 columns

Code

open_streets_days = join_test[['appronstre', 'BoroName', 'monday', 'tuesday', 'wednesday','thursday', 'friday', 'saturday', 'sunday']]
open_streets_days

	appronstre	BoroName	monday	tuesday	wednesday	thursday	friday	saturday	sunday
0	DEISIUS STREET	Staten Island	1	1	1	1	1	0	0
1	SUFFOLK AVENUE	Staten Island	0	0	0	0	1	1	1
2	SUFFOLK AVENUE	Staten Island	0	0	0	0	1	1	1
3	VERMONT COURT	Staten Island	0	0	0	0	1	1	1
4	9 STREET	Staten Island	0	0	0	0	1	1	1
...	...	...	...	...	...	...	...	...	...
361	BECK STREET	Bronx	0	0	1	0	0	0	0
362	NEWKIRK AVENUE	Brooklyn	0	0	0	0	0	0	1
363	18 STREET	Brooklyn	1	1	1	1	1	0	0
364	34 AVENUE	Queens	1	1	1	1	1	1	1
365	DECATUR STREET	Brooklyn	0	0	0	0	0	1	0

366 rows × 9 columns

Census Data

Code

Pop_2020 = gpd.read_file("data/NYC_tracts_2020.geojson")
Pop_2020 = Pop_2020[['GEOID', 'estimate']]
Pop_2020['GEOID']=Pop_2020['GEOID'].astype(int)

tracts_pop = tracts_clean.merge(Pop_2020, on='GEOID', how='left') 
neighbor_pop = tracts_pop.groupby(['NTAName']).sum(['estimate']).reset_index()

Code

income = gpd.read_file("data/NYC_Income.geojson")
geo_income = gpd.GeoDataFrame(income, geometry='geometry')
geo_income = geo_income.set_crs(epsg=4326)
geo_income['GEOID']=geo_income['GEOID'].astype('int')

tracts_income = tracts_clean.merge(geo_income.drop(columns='geometry'), on='GEOID', how='left').dropna()
neighbor_income = tracts_income.groupby(['NTAName', 'BoroName']).median(['estimate']).reset_index()

Chart I: Matplotlib – Parks in Neighborhoods

I first hope to investigate into the distribution of Parks in different neighborhoods and boroughs in relation to population. On the one hand, higher population means more people will have need for a bigger public green space for leisure time. On the other hand, less populated neighborhoods tend to have more spaces for parks. And since the count of parks doesn’t perfectly reflect how much space is available, I am using acrage data instead of counts for this analysis.

To investigate, I utilized Matplotlib, which is great for making simple scatterplot charts that speaks for simple linear relationship, if there is any.

Code

geo_parks_clean['acres'] = geo_parks_clean['acres'].astype(float)
parks_acres_neighborhood = geo_parks_clean.groupby('NTAName').sum().drop(columns=['index_right','CT2020']).reset_index()
parks_acres_pop = neighbor_pop.merge(parks_acres_neighborhood, on='NTAName', how='left').dropna()
parks_pop = parks_acres_pop.merge(Boro_NTA, on='NTAName', how='left').dropna()
parks_pop['acres'].describe()

/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1268464147.py:2: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.
  parks_acres_neighborhood = geo_parks_clean.groupby('NTAName').sum().drop(columns=['index_right','CT2020']).reset_index()

count     237.000000
mean      105.945812
std       224.643587
min         0.005000
25%         7.651000
50%        20.653000
75%        85.527000
max      1930.636136
Name: acres, dtype: float64

After a quick glance of the park data, I notice some extreme outliers with extremely large parks that not only serves adjacent neighborhoods but the whole city. I excluded those outliers to have a better sense of how much acrage of parks common neighborhoods get.

Code

parks_pop_filtered = parks_pop.loc[(parks_pop['acres'] < 86) & (parks_pop['acres'] > 7) & (parks_pop['estimate'] >0)]
parks_pop_filtered

	NTAName	CT2020	GEOID_x	estimate	acres	GEOID_y	BoroName
5	Astoria (East)-Woodside (North)	241600	505134241600	34825.0	8.642	2.886482e+11	Queens
6	Astoria (North)-Ditmars-Steinway	214102	613377214102	47134.0	11.530	4.690532e+11	Queens
10	Barren Island-Floyd Bennett Field	70202	36047070202	26.0	64.665	3.604707e+10	Brooklyn
11	Bath Beach	235800	396517235800	32716.0	21.398	1.081411e+11	Brooklyn
15	Bedford Park	448115	396055448115	55702.0	23.917	1.800252e+11	Bronx
...	...	...	...	...	...	...	...
225	West Farms	136300	180025136300	18206.0	13.540	3.960554e+11	Bronx
230	Whitestone-Beechhurst	703300	252567703300	28353.0	29.086	2.164866e+11	Queens
232	Williamsburg	695100	468611695100	59410.0	81.282	1.045365e+12	Brooklyn
233	Windsor Terrace-South Slope	363808	288376363808	25442.0	8.408	4.325650e+11	Brooklyn
235	Woodside	406004	505134406004	45417.0	10.217	5.412154e+11	Queens

113 rows × 7 columns

Code

color_map = {"Bronx": "#550527", "Brooklyn": "#688E26", "Manhattan": "#FAA613", "Queens": "#F44708", "Staten Island": "#A10702"}

fig, ax = plt.subplots(figsize=(11,6))

for BoroName, group_df in parks_pop_filtered.groupby("BoroName"):
    
    ax.scatter(
        group_df["estimate"],
        group_df["acres"],
        marker="P",
        label=BoroName,
        color=color_map[BoroName],
        alpha=0.75,
        zorder=10
    )

ax.legend(loc="best")
ax.set(
    title = "Park Space (acres) Relative to Population in Neighborhoods, by Boroughs",
    xlabel = "Population in each neighborhood (2020 Census)",
    ylabel = "Parks in Acres")
ax.grid(True)

plt.show()

This chart does not suggest a strong linear relationship between population and park acrages. Overall, Manhattan, Queens and Brooklyn host more large parks, but some neighborhoods are particularly underserved, with very high population and low park acrage (points towards lower right of the chart).

Chart II: Seaborn (x2)

Looking at data for indoor leisure space and open street data, I utilize seaborn to create two types of charts that suite the nature of the data sets.

The first is a grouped bar charts. Similar to park distribution, I hope to have a quick glimpse of number of each type of leisure spaces in each borough, and hope to identify any general spatial patterns or inequalities. A grouped chart is great a revealing such pattern.

The second is a heatmap for open street data. The heatmap explores the number of open streets approved in different boroughs on different days of the week.

Indoor Leisure Spaces: Grouped Bar Charts

Code

indoor_clean = geo_total_indoor.groupby(['Type','BoroName']).count().reset_index().drop(columns=['Zip','Address','geometry','index_right','CT2020','NTAName']).pivot(index='Type',columns='BoroName', values='Name').fillna(0).reset_index()
indoor_melt = indoor_clean.melt(id_vars='Type', value_vars=['Bronx','Brooklyn','Manhattan', 'Queens', 'Staten Island'])
indoor_melt

	Type	BoroName	value
0	Art Galleries	Bronx	6.0
1	Libraries	Bronx	35.0
2	Museums	Bronx	8.0
3	Theatres	Bronx	0.0
4	Art Galleries	Brooklyn	61.0
5	Libraries	Brooklyn	59.0
6	Museums	Brooklyn	12.0
7	Theatres	Brooklyn	0.0
8	Art Galleries	Manhattan	823.0
9	Libraries	Manhattan	44.0
10	Museums	Manhattan	87.0
11	Theatres	Manhattan	115.0
12	Art Galleries	Queens	24.0
13	Libraries	Queens	65.0
14	Museums	Queens	12.0
15	Theatres	Queens	2.0
16	Art Galleries	Staten Island	3.0
17	Libraries	Staten Island	13.0
18	Museums	Staten Island	9.0
19	Theatres	Staten Island	0.0

Code

sns.set_theme(style="whitegrid")

color_map = ["#550527", "#688E26", "#FAA613", "#F44708", "#A10702"]
sns.set_palette(color_map)

sns.catplot(
    data=indoor_melt, kind="bar",
    x="Type", 
    y="value", 
    hue="BoroName",
    aspect=2, 
    alpha=1
).set_axis_labels(
    "Type of Leisure Space", "Counts"
).set(title="Distribution of 4 Types of Leisure Spaces in Each Borough")

/Users/annzhang/mambaforge/envs/musa-550-fall-2023/lib/python3.10/site-packages/seaborn/axisgrid.py:118: UserWarning: The figure layout has changed to tight
  self._figure.tight_layout(*args, **kwargs)

Days and Location of Open Streets in NYC: A Heatmap

Code

#heatmap for open_streets: borough x Days of the week 
open_streets_days_melt= open_streets_days.melt(id_vars=['appronstre','BoroName'], value_vars=['monday','tuesday','wednesday', 'thursday', 'friday','saturday','sunday'])

	appronstre	BoroName	variable	value
0	DEISIUS STREET	Staten Island	monday	1
1	SUFFOLK AVENUE	Staten Island	monday	0
2	SUFFOLK AVENUE	Staten Island	monday	0
3	VERMONT COURT	Staten Island	monday	0
4	9 STREET	Staten Island	monday	0
...	...	...	...	...
2557	BECK STREET	Bronx	sunday	0
2558	NEWKIRK AVENUE	Brooklyn	sunday	1
2559	18 STREET	Brooklyn	sunday	0
2560	34 AVENUE	Queens	sunday	1
2561	DECATUR STREET	Brooklyn	sunday	0

2562 rows × 4 columns

Code

open_streets_seaborn = open_streets_days_melt.groupby(['variable','BoroName']).sum().reset_index()

# sort the order of day from monday to sunday 

order=['monday', 'tuesday', 'wednesday', 'thursday', 'friday', 'saturday', 'sunday']
open_streets_seaborn['variable'] = pd.Categorical(open_streets_seaborn['variable'], categories=order, ordered=True)
open_streets_seaborn = open_streets_seaborn.sort_values(by='variable')

/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/3045256453.py:1: FutureWarning: The default value of numeric_only in DataFrameGroupBy.sum is deprecated. In a future version, numeric_only will default to False. Either specify numeric_only or select only columns which should be valid for the function.
  open_streets_seaborn = open_streets_days_melt.groupby(['variable','BoroName']).sum().reset_index()

Code

import seaborn as sns

sns.set_theme()

# Load the example flights dataset and convert to long-form

open_street_heatmap = (
    open_streets_seaborn
    .pivot(index="BoroName", columns="variable", values="value")
)

# Draw a heatmap with the numeric values in each cell
f, ax = plt.subplots(figsize=(9, 6))

ax = sns.heatmap(open_street_heatmap, annot=True, linewidths=.5)
ax.set(xlabel="Days in a week", ylabel="Borough", title="Number of Streets Open on Specific Days")

[Text(0.5, 33.249999999999986, 'Days in a week'),
 Text(79.75, 0.5, 'Borough'),
 Text(0.5, 1.0, 'Number of Streets Open on Specific Days')]

From both charts, we can see Manhattan has disproportional number of art galleries and open streets in comparison to other boroughs. Museums and theatres are predominantly located on Manhattan as well. Interestingly, the distribution of libraries seems more even. However, overall, we are seeing leisure spaces are disproportional abundant and diverse in Manhattan, then brooklyn and queens, leaving Bronx and Staten Island less resourceful.

Chart III: Altair Charts (x3)

Furthering the exploration on indoor leisure spaces, I created one chart on its ditribution in relation with median household income, to see if it embeds any socio-economic inequality. I then creatd a map to visualize this relationship, which can also be helpful to locate different kinds of leisure spaces in the city.

My third chart is an interactive bar chart on open streets in Brookylen, where audience can choose the day of the week to see all open streets approved for that day and the time they open and close. This could potentially be developed into a tool for residents and tourists to track open streets.

Brush Selection: Indoor Space and Income

Code

neighbor_indoor = geo_total_indoor.groupby(['NTAName','Type']).count().reset_index().drop(['Zip', 'Address', 'geometry','index_right','BoroName', 'CT2020'], axis=1)
neighbor_indoor = neighbor_indoor.rename(columns={'Name': 'Count'})
neighbor_income_indoor = neighbor_income.merge(neighbor_indoor, on='NTAName', how='left')

neighbor_income_indoor_pivot=neighbor_income_indoor.pivot(index=['NTAName', 'BoroName','estimate'], columns="Type", values="Count").reset_index().fillna(0)
neighbor_income_indoor_pivot_1 = neighbor_income_indoor_pivot.drop(neighbor_income_indoor_pivot.columns[[3]],axis=1)
income_indoor_pivot = neighbor_income_indoor_pivot_1.melt(id_vars=["NTAName", "estimate", "BoroName"], value_vars=["Art Galleries", "Libraries", "Museums", "Theatres"], var_name="Types",value_name="Count")
income_indoor_pivot_filtered = income_indoor_pivot.loc[(income_indoor_pivot['Count'] <20)]
income_indoor_pivot_filtered = income_indoor_pivot_filtered.rename(columns={'estimate': 'Median Household Income'})

Code

brush = alt.selection_interval()

Brush_Chart = (
alt.Chart(income_indoor_pivot_filtered)
   .mark_point()
   .encode(
       x=alt.X("Median Household Income:Q", scale=alt.Scale(zero=False)),
       y=alt.Y("Count:Q", scale=alt.Scale(zero=False)),
       color=alt.condition(brush, "BoroName:N", alt.value("lightgray")),
       tooltip=["NTAName","BoroName:N", "Median Household Income:Q", "Count:Q"])
   .add_params(brush)
   .properties(width=200, height=200)
   .facet(column="Types:N")
)

Brush_Chart

Map: Income and Indoor Leisure Spaces - Relationships

Code

NTA = pd.read_csv("data/2020 Neighborhood.csv")
NTA['geometry'] = gpd.GeoSeries.from_wkt(NTA['the_geom'])
NTA_geo = gpd.GeoDataFrame(NTA, geometry='geometry', crs=4326)
NTA_geo = NTA_geo.to_crs(epsg=2263)
geo_total_indoor = geo_total_indoor.to_crs(epsg=2263)

geo_NTA = NTA_geo[['NTAName', 'geometry']]

tracts_income_1 = tracts_clean.merge(geo_income.drop(columns='geometry'), on='GEOID', how='left').dropna()
neighbor_income_1 = tracts_income.groupby(['NTAName', 'BoroName']).median(['estimate']).reset_index()

NTA_income = geo_NTA.merge(neighbor_income_1, on='NTAName', how='left')

Code

geo_total_indoor_1 = geo_total_indoor

geo_total_indoor_1['lon'] = geo_total_indoor_1['geometry'].x
geo_total_indoor_1['lat'] = geo_total_indoor_1['geometry'].y

Code

Income = (
    alt.Chart(NTA_income)
    .mark_geoshape(stroke="white")
    .encode(
        tooltip=["NTAName:N", "estimate:Q", "moe:Q"],
        color=alt.Color("estimate:Q", scale=alt.Scale(scheme="greys")),
    )
    # Important! Otherwise altair will try to re-project your data
    .project(type="identity", reflectY=True)
    .properties(width=1000, height=800).interactive()
)

IndoorSpaces = (
    alt.Chart(geo_total_indoor_1)
    .mark_circle(size=10)
    .encode(tooltip=['Name','Type','Address'],
           longitude="lon", latitude="lat",
           color=alt.Color('Type:N', scale=alt.Scale(scheme="lightmulti"))
         ).project(type="identity", reflectY=True)
)



map_1 = Income + IndoorSpaces
map_1

Similarly, libraries seem to be the least discriminatory type of leisure space. For art galleries and museums, while there are some distributed in mid to lower income neighborhood, the higher income neighborhood has higher density of such leisure space. Theatres, at the same time, is mostly located in mid to higher income neighborhood. This trend can be clearly see on the map too. Such spatial distribution means, for residents in lower to mid income neighborhood, they might have to travel further for accessing those spaces.

Open Street Time

Code

open_csv = pd.read_csv("data/Open Streets CSV.csv")
open_csv['geometry'] = gpd.GeoSeries.from_wkt(open_csv['the Geom'])
open_csv_geo = gpd.GeoDataFrame(open_csv, geometry='geometry', crs=2263)
open_csv_geo = open_csv_geo.to_crs(epsg=4326)

open_7days_time = open_csv_geo.drop(['apprDaysWe','Object ID', 'Organization Name', 'Approved From Street', 'Approved To Street', 'apprStartD', 'apprEndDat', 'Shape_STLe', 'segmentidt', 'segmentidf', 'lionversion', 'the Geom'], axis=1)
open_7days_time = open_7days_time.drop_duplicates(subset = "Approved On Street")
open_7days_time_melt = open_7days_time.melt(id_vars=['Approved On Street', 'Borough Name'], value_vars=['Approved Monday Open', 'Approved Monday Close', 'Approved Tuesday Open', 'Approved Tuesday Close', 'Approved Wednesday Open', 'Approved Wednesday Close', 'Approved Thursday Open', 'Approved Thursday Close', 'Approved Friday Open', 'Approved Friday Close', 'Approved Saturday Open', 'Approved Saturday Close', 'Approved Sunday Open', 'Approved Sunday Close']).dropna()
#open_7days_time_melt['value'] =  pd.to_datetime(open_7days_time_melt['value']).dt.time
open_time_Brooklyn = open_7days_time_melt[(open_7days_time_melt['Borough Name'] == 'Brooklyn')]
open_time_Brooklyn

	Approved On Street	Borough Name	variable	value
6	RIDGE BOULEVARD	Brooklyn	Approved Monday Open	10:00
9	82 STREET	Brooklyn	Approved Monday Open	08:30
10	48 STREET	Brooklyn	Approved Monday Open	13:30
11	43 STREET	Brooklyn	Approved Monday Open	09:30
12	ALBEMARLE ROAD	Brooklyn	Approved Monday Open	08:00
...	...	...	...	...
2282	JEFFERSON AVENUE	Brooklyn	Approved Sunday Close	21:00
2287	SHARON STREET	Brooklyn	Approved Sunday Close	20:00
2288	TROUTMAN STREET	Brooklyn	Approved Sunday Close	22:00
2289	RANDOLPH STREET	Brooklyn	Approved Sunday Close	23:00
2354	LEXINGTON AVENUE	Brooklyn	Approved Sunday Close	20:00

378 rows × 4 columns

Code

open_time_Brooklyn['time'] = open_7days_time_melt['variable'].str.extract('(Open|Close)')
open_time_Brooklyn['dayweek'] = open_7days_time_melt['variable'].str.extract('(Monday|Tuesday|Wednesday|Thursday|Friday|Saturday|Sunday)')
open_time_Brooklyn= open_time_Brooklyn.pivot(index=['Approved On Street', 'dayweek'], columns="time", values="value").reset_index()

order=['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday']
open_time_Brooklyn['dayweek'] = pd.Categorical(open_time_Brooklyn['dayweek'], categories=order, ordered=True)
open_time_Brooklyn = open_time_Brooklyn.sort_values(by='dayweek').reset_index()

open_time_Brooklyn

/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1900940832.py:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  open_time_Brooklyn['time'] = open_7days_time_melt['variable'].str.extract('(Open|Close)')
/var/folders/q3/y0zpvj752qg3_3nvpkx6v2300000gn/T/ipykernel_80527/1900940832.py:2: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  open_time_Brooklyn['dayweek'] = open_7days_time_melt['variable'].str.extract('(Monday|Tuesday|Wednesday|Thursday|Friday|Saturday|Sunday)')

time	index	Approved On Street	dayweek	Close	Open
0	139	SUMMIT STREET	Monday	14:30	12:30
1	151	UNDERHILL AVENUE	Monday	20:00	08:00
2	28	82 STREET	Monday	15:30	08:30
3	143	THATFORD AVENUE	Monday	18:00	11:00
4	33	AITKEN PLACE	Monday	15:00	11:00
...	...	...	...	...	...
184	147	TOMPKINS AVENUE	Sunday	20:00	11:00
185	51	BEVERLEY ROAD	Sunday	18:00	10:00
186	112	RANDOLPH STREET	Sunday	23:00	12:00
187	10	4 STREET	Sunday	22:00	08:00
188	157	VANDERBILT AVENUE	Sunday	23:00	11:00

189 rows × 5 columns

Code

selection = alt.selection_multi(fields=['dayweek'])
color = alt.condition(selection,
                      alt.Color('dayweek:N', legend=None, 
                      scale=alt.Scale(scheme='category10')),
                      alt.value('lightgray'))

opacity = alt.condition(selection,
                        alt.value(1), alt.value(0))



bar = alt.Chart(open_time_Brooklyn).mark_bar().encode(
        x='Open',
        x2='Close',
        y='Approved On Street',
        color=color,
        opacity=opacity,
        tooltip=['Open', 'Close', 'Approved On Street', 'dayweek']).properties(
        width=500,
        height=1000).interactive()

legend = alt.Chart(open_time_Brooklyn).mark_bar().encode(
    y=alt.Y('dayweek:N', axis=alt.Axis(orient='right')),
    color=color
).add_selection(
selection
)

A_Chart = bar | legend
A_Chart

This interactive bar chart on Brooklyn Open Street serves as a pilot that can be adapted for data of all five boroughs. By comparing to open and close time, we can observe that many streets are approved to be open streets with later opening and closing time on weekends. To improve this bar chart the status of each street can be added (i.e., whether they are approved to be closed fully or partially or only on school days), which is important for visitors as well.

Dashboard

The dashboard below is an upgrade from the indoor space and income grouped scatter plot. This dashboard makes it easier to explore the spatial distribution of indoor spaces of certain type or in neighborhoods with certain level of income. Through selecting points with higher counts, we can see neighborhoods with higher density of identified indoor leisure spaces are mostly located in Manhattan.

Code

brush = alt.selection_interval()

points = alt.Chart(income_indoor_pivot_filtered).mark_point().encode(
       x=alt.X("Median Household Income:Q", scale=alt.Scale(zero=False)),
       y=alt.Y("Count:Q", scale=alt.Scale(zero=False)),
       color=alt.condition(brush, "BoroName:N", alt.value("lightgray")),
       tooltip=["NTAName","BoroName:N", "Median Household Income:Q", "Count:Q"]).add_params(brush
).properties(width=200, height=200
).facet(column="Types:N")


bars = alt.Chart(income_indoor_pivot_filtered).mark_bar().encode(
    y='BoroName:N',
    color='BoroName:N',
    x='Count:Q'
).transform_filter(
    brush
)

Dashboard = points & bars
Dashboard