Exploring Social Capital

blog

policy-impact

data-analysis

place-based-change

Exploration of novel Social Connections data to support Lloyds Bank Foundation with their place-based change work

Author

Just Knowledge Team

Published

October 17, 2025

Exploring Social Capital

Jolyon Miles-Wilson

17-10-2025

We’re currently exploring some opportunities to support Lloyds Bank Foundation with their place-based change work using novel data on Social Capital, based on data from Facebook.¹ The aim is to understand how social capital can help identify and shape development stategies for areas at a hyperlocal level.

The Data

The dataset is based on 20.5 million active UK Facebook users with at least 100 friends, aged 25-64, which represents approximately 58% of the 25-64 UK population.² It was produced through a collaboration between Meta, Behavioural Insights Team, the Royal Society of Arts, Stripe Partners, Neighbourly Lab, Opportunity Insights, New York University and Stanford University. During our time at the RSA, both Celestin and I were closely involved in this work.

The data contain many variables measuring different facets of social capital. For now, we focus on just one: Economic Connectedness. Economic Connectedness represents the proportion of friends of people of low socioeconomic status (SES) who are of high SES. In the seminal work in the US context by Chetty et al. (2022),³ Economic Connectedness was found to strongly predict upward social mobility.

The National View

Social capital is stronger in the UK than it is in the US. But it is qualitatively different; in the UK it is driven by hobby groups, whereas in the US it is driven by religious groups.

While the overall picture is strong, social capital varies across England and Wales.

The Regional Picture

Social Capital varies from place to place, which may make it a particularly useful metric for judging the strengths and needs of local areas. Below, we can see how social capital varies between and within regions.⁴

This map shows Economic Connectedness by Local Authority District for the North East.

Purple values indicate lower Economic Connectedness.

Whiter values indicate higher Economic Connectedness.

This map shows Economic Connectedness by Local Authority District for the West Midlands.

Purple values indicate lower Economic Connectedness.

Whiter values indicate higher Economic Connectedness.

This map shows Economic Connectedness by Local Authority District for the East of England.

Purple values indicate lower Economic Connectedness.

Whiter values indicate higher Economic Connectedness.

The Work continues

We aim to further explore how these novel data on social capital can be put to use to help improve social outcomes. Stay tuned by subscribing to our newsletter below to hear about our progress.

Footnotes

https://data.humdata.org/dataset/uk-social-capital-atlas ↩︎
Harris, T., Iyer, S., Rutter, T., Chi, G., Johnston, D., Lam, P., Makinson, L., Silva, A., Wessel, M., Liou, M.-C., Wang, Y., Zaman, Q., & Bailey, M. (2025). Social Capital in the United Kingdom: Evidence from Six Billion Friendships. OSF. https://doi.org/10.31235/osf.io/kb7dy_v1 ↩︎
Chetty, R., Jackson, M. O., Kuchler, T., Stroebel, J., Hendren, N., Fluegge, R. B., Gong, S., Gonzalez, F., Grondin, A., Jacob, M., Johnston, D., Koenen, M., Laguna-Muggenburg, E., Mudekereza, F., Rutter, T., Thor, N., Townsend, W., Zhang, R., Bailey, M., … Wernerfelt, N. (2022). Social capital I: Measurement and associations with economic mobility. Nature, 608(7921), 108–121. https://doi.org/10.1038/s41586-022-04996-4 ↩︎
We focus here only on this handful of areas because these are the places where some of Lloyds Bank Foundation’s work is already focused. The same exercise can be done for any place in the UK↩︎

--- title: "Exploring Social Capital" description: "Exploration of novel Social Connections data to support Lloyds Bank Foundation with their place-based change work" date: "2025-10-17" categories: [policy-impact, data-analysis, place-based-change] echo: false draft: false image: "/assets/images/blog/social-connections.png" code-tools: true code-summary: "Code For Nerds" --- ![](../assets/images/blog/social-connections.png){fig-align="center" height=300} # Exploring Social Capital ::: {.columns} ::: {.column width="50%"} <div style="text-align: center;"> *Jolyon Miles-Wilson* </div> ::: ::: {.column width="50%"} <div style="text-align: center;"> *17-10-2025* </div> ::: :::  We're currently exploring some opportunities to support Lloyds Bank Foundation with their **place-based change** work using novel data on Social Capital, based on data from Facebook.^[[https://data.humdata.org/dataset/uk-social-capital-atlas](https://data.humdata.org/dataset/ uk-social-capital-atlas)] The aim is to understand how social capital can help identify and shape development stategies for areas at a hyperlocal level. ## The Data The dataset is based on 20.5 million active UK Facebook users with at least 100 friends, aged 25-64, which represents approximately 58% of the 25-64 UK population.^[Harris, T., Iyer, S., Rutter, T., Chi, G., Johnston, D., Lam, P., Makinson, L., Silva, A., Wessel, M., Liou, M.-C., Wang, Y., Zaman, Q., & Bailey, M. (2025). Social Capital in the United Kingdom: Evidence from Six Billion Friendships. OSF. [https://doi.org/10.31235/osf.io/kb7dy_v1](https://doi.org/10.31235/osf.io/kb7dy_v1)] It was produced through a collaboration between Meta, Behavioural Insights Team, the Royal Society of Arts, Stripe Partners, Neighbourly Lab, Opportunity Insights, New York University and Stanford University. During our time at the RSA, both Celestin and I were closely involved in this work. The data contain many variables measuring different facets of social capital. For now, we focus on just one: **Economic Connectedness**. Economic Connectedness represents the **proportion of friends of people of low socioeconomic status (SES) who are of high SES**. In the seminal work in the US context by Chetty et al. (2022),^[Chetty, R., Jackson, M. O., Kuchler, T., Stroebel, J., Hendren, N., Fluegge, R. B., Gong, S., Gonzalez, F., Grondin, A., Jacob, M., Johnston, D., Koenen, M., Laguna-Muggenburg, E., Mudekereza, F., Rutter, T., Thor, N., Townsend, W., Zhang, R., Bailey, M., … Wernerfelt, N. (2022). Social capital I: Measurement and associations with economic mobility. Nature, 608(7921), 108–121. [https://doi.org/10.1038/s41586-022-04996-4](https://doi.org/10.1038/s41586-022-04996-4)] Economic Connectedness was found to strongly predict upward social mobility. ```{python} # Packages and config import pandas as pd import requests import re import geopandas as gpd from shapely.geometry import Point import matplotlib.pyplot as plt import matplotlib as mpl from mpl_toolkits.axes_grid1.axes_divider import make_axes_locatable import contextily as cx import configparser import janitor import zipfile import glob from sqlalchemy import create_engine import matplotlib.colors as mcolors import pickle import os import numpy as np config = configparser.ConfigParser() config.read(os.path.join('..', 'db_config.ini')) db_params = dict(config['postgresql']) # Load config config = configparser.ConfigParser() config.read(os.path.join('..', 'db_config.ini')) db_params = dict(config['postgresql']) # Build SQLAlchemy connection string conn_str = ( f"postgresql+psycopg2://{db_params['user']}:{db_params['password']}" f"@{db_params['host']}:{db_params['port']}/{db_params['database']}" ) # Create engine engine = create_engine(conn_str) ``` ```{python} # Updated to appropriate file path filepath = os.path.join('..', 'assets', 'palettes', 'ice_swatch.txt') with open(filepath, 'r', encoding='utf-8') as file: ice_swatch = [line.strip() for line in file if line.strip()] # Make a colormap ice_cmap = mcolors.LinearSegmentedColormap.from_list("ice", ice_swatch).reversed() # Updated to appropriate file path with open(os.path.join('..', 'assets', 'palettes', 'jk_primary_colours.txt'), 'r') as file: jk_colours = [line.strip() for line in file] # Font hfont = {'fontname': 'Inclusive Sans'} nfont = {'fontname': 'Open Sans'} # Set global font family and size plt.rcParams['font.family'] = 'Open Sans' plt.rcParams['axes.titlesize'] = 14 # size of the title plt.rcParams['axes.labelsize'] = 12 mpl.rcParams["figure.facecolor"] = jk_colours[0] ``` ```{python} # Download data if it doesn't exist, otherwise load it path = os.path.join('data') if not os.path.exists(path): os.makedirs(path) # LAD file = 'lad_gpd.pkl' filepath = path + '/' + file if not os.path.exists(filepath): with engine.connect() as con: query = ''' WITH lookup AS ( SELECT DISTINCT ON (lad21cd) loo.lad21cd, loo.ladnm as lad21nm, foo.rgn21nm_filled as rgn21nm FROM pcode_census21_lookup loo LEFT JOIN lad21_lookup foo ON loo.lad21cd = foo.lad21cd ) SELECT lookup.*, geom.geometry FROM lookup LEFT JOIN lad21_boundaries geom ON lookup.lad21cd = geom.lad21cd; ''' lad_gpd = gpd.read_postgis(query, con, geom_col='geometry') with open(filepath, 'wb') as file: pickle.dump(lad_gpd, file) else: with open(filepath, 'rb') as file: lad_gpd = pickle.load(file) # MSOA file = 'msoa_gpd.pkl' filepath = path + '/' + file if not os.path.exists(filepath): with engine.connect() as con: query = ''' WITH lookup AS ( SELECT DISTINCT ON (msoa21cd) loo.msoa21cd, loo.lad21cd, loo.ladnm as lad21nm, foo.rgn21nm_filled as rgn21nm FROM pcode_census21_lookup loo LEFT JOIN lad21_lookup foo ON loo.lad21cd = foo.lad21cd ) SELECT lookup.*, geom.geometry FROM lookup LEFT JOIN msoa21_boundaries geom ON lookup.msoa21cd = geom.msoa21cd; ''' msoa_gpd = gpd.read_postgis(query, con, geom_col='geometry') with open(filepath, 'wb') as file: pickle.dump(msoa_gpd, file) else: with open(filepath, 'rb') as file: msoa_gpd = pickle.load(file) # MSOA 11 file = 'msoa11_gpd.pkl' filepath = path + '/' + file if not os.path.exists(filepath): with engine.connect() as con: query = ''' WITH lookup AS ( SELECT DISTINCT ON (msoa11cd) loo.msoa11cd, loo.lad21cd, loo.lad21nm, foo.rgn21nm_filled as rgn21nm FROM pcode_census11_lookup loo LEFT JOIN lad21_lookup foo ON loo.lad21cd = foo.lad21cd ) SELECT lookup.*, geom.geometry FROM lookup LEFT JOIN msoa11_boundaries geom ON lookup.msoa11cd = geom.msoa11cd; ''' msoa11_gpd = gpd.read_postgis(query, con, geom_col='geometry') with open(filepath, 'wb') as file: pickle.dump(msoa11_gpd, file) else: with open(filepath, 'rb') as file: msoa11_gpd = pickle.load(file) file = 'lookup.pkl' filepath = path + '/' + file if not os.path.exists(filepath): with engine.connect() as con: query = ''' SELECT DISTINCT ON (msoa21cd) loo.msoa21cd, loo.lad21cd, loo.ladnm as lad21nm, foo.rgn21nm_filled as rgn21nm FROM pcode_census21_lookup loo LEFT JOIN lad21_lookup foo ON loo.lad21cd = foo.lad21cd ''' lookup = pd.read_sql(sql=query, con=con) with open(filepath, 'wb') as file: pickle.dump(lookup, file) else: with open(filepath, 'rb') as file: lookup = pickle.load(file) ``` ```{python} #| output: false #| # Get the actual published data ### LAD ### # 1. Specify the url that downloads the data url = 'https://data.humdata.org/dataset/79d5959f-73b8-4305-bb9e-d829e2e863da/resource/f87c1d44-a56c-48cf-a8bb-1caad97aa589/download/local_authority-20250320t000233z-001.zip' # 2. Specify folder name as the final part of the url after it's been split on '/' and specify relative path folder = url.split('/')[-1] path = os.path.join('data', folder) # 3. If the path doesn't exist, download the data. If it does exist, skip this step. if not os.path.exists(path): req = requests.get(url) with open(path, 'wb') as output_file: output_file.write(req.content) else: print('Data already downloaded. Loading') # 4. Unzip the folder ## i. Define outpath as same as in path minus .zip out_path = os.path.splitext(path)[0] # This removes the .zip extension ## ii. Create the extraction directory if it doesn't exist if not os.path.exists(out_path): os.makedirs(out_path) ## iii. Unzip to extraction directory with zipfile.ZipFile(path, 'r') as zip_ref: zip_ref.extractall(out_path) # 5. Identify the file in the directory you want to read and read it filepath = glob.glob( os.path.join(out_path, "local_authority", "*_economic_connectedness_[0-9][0-9][0-9][0-9]*.csv") ) meta_lad = pd.read_csv(filepath[0]) meta_lad = meta_lad.drop(meta_lad.columns[0], axis=1) # For now let's concentrate only on low_ses_users. But we might # be interested to look at high_ses_users or the ratio of EC # between low and high users # We do this here so that we retain all areas in the subsequent join. We want to ensure # we have all areas and assign NA to areas not in the Meta data meta_lad = meta_lad.loc[meta_lad['measurement_category']=='low_ses_users']\ .rename(columns= {'high_ses_ratio': 'economic_connectedness'} ) meta_lad = lad_gpd.merge(meta_lad, how='left', left_on='lad21cd', right_on='lad_code').drop(columns='lad_code') # Make decile variable meta_lad['ec_decile'] = pd.qcut(meta_lad['economic_connectedness'], q=10, labels=[x for x in range(10,0,-1)], duplicates="drop") ### MSOA #### # 1. Specify the url that downloads the data url = 'https://data.humdata.org/dataset/79d5959f-73b8-4305-bb9e-d829e2e863da/resource/b1c4a576-c469-4447-b822-7215a39afcd8/download/msoa-20250320t000309z-001.zip' # 2. Specify folder name as the final part of the url after it's been split on '/' and specify relative path folder = url.split('/')[-1] path = os.path.join('data', folder) # 3. If the path doesn't exist, download the data. If it does exist, skip this step. if not os.path.exists(path): req = requests.get(url) with open(path, 'wb') as output_file: output_file.write(req.content) else: print('Data already downloaded. Loading') # 4. Unzip the folder ## i. Define outpath as same as in path minus .zip out_path = os.path.splitext(path)[0] # This removes the .zip extension ## ii. Create the extraction directory if it doesn't exist if not os.path.exists(out_path): os.makedirs(out_path) ## iii. Unzip to extraction directory with zipfile.ZipFile(path, 'r') as zip_ref: zip_ref.extractall(out_path) # 5. Identify the file in the directory you want to read and read it filepath = glob.glob( os.path.join(out_path, "msoa", "*_economic_connectedness_[0-9][0-9][0-9][0-9]*.csv") ) meta_msoa = pd.read_csv(filepath[0]) meta_msoa = meta_msoa.drop(meta_msoa.columns[0], axis=1) # For now let's concentrate only on low_ses_users. But we might # be interested to look at high_ses_users or the ratio of EC # between low and high users # We do this here so that we retain all areas in the subsequent join. We want to ensure # we have all areas and assign NA to areas not in the Meta data meta_msoa = meta_msoa.loc[meta_msoa['measurement_category']=='low_ses_users']\ .rename(columns= {'high_ses_ratio': 'economic_connectedness'} ) meta_msoa = msoa_gpd.merge(meta_msoa, how='left', left_on='msoa21cd', right_on='zip_prediction').drop(columns='zip_prediction') # Make decile variable meta_msoa['ec_decile'] = pd.qcut(meta_msoa['economic_connectedness'], q=10, labels=[x for x in range(10,0,-1)], duplicates="drop") ``` ## The National View Social capital is stronger in the UK than it is in the US. But it is qualitatively different; in the UK it is driven by hobby groups, whereas in the US it is driven by religious groups. While the overall picture is strong, social capital varies across England and Wales. ```{python} #| fig-align: center # variable = 'economic_connectedness' variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') fig, ax = plt.subplots(1,1, figsize = [15,10]) meta_lad.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.8, legend=False, edgecolor='black', linewidth=0.01, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True } ) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) cx.add_basemap(ax, crs=meta_lad.crs, source=cx.providers.CartoDB.Positron) # ax.set_title(f"{variable_clean}: England and Wales", fontsize=14, **hfont) ax.set_axis_off(); ```  ```{python} lads = ['Great Yarmouth','Redcar and Cleveland','Telford and Wrekin'] areas_of_interest = lookup.loc[lookup['lad21nm'].isin(lads),['lad21nm','rgn21nm']].drop_duplicates().reset_index(drop=True) ``` ## The Regional Picture {.smaller} Social Capital varies from place to place, which may make it a particularly useful metric for judging the strengths and needs of local areas. Below, we can see how social capital varies between and within regions.^[We focus here only on this handful of areas because these are the places where some of Lloyds Bank Foundation's work is already focused. The same exercise can be done for any place in the UK] ::: {.panel-tabset} ```{python} area_of_interest = areas_of_interest.iloc[0,0] region_of_interest = areas_of_interest.iloc[0,1] ``` ### `{python} region_of_interest` This map shows Economic Connectedness by **Local Authority District** for the **`{python} region_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') meta_subset = meta_lad.loc[meta_lad['rgn21nm'].str.contains(region_of_interest, na=False)] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.1, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) aoi = meta_subset.loc[meta_subset['lad21nm']==area_of_interest] xy = np.array([aoi.centroid.x.values[0], aoi.centroid.y.values[0]]) nudge = [1, .97] xy_nudged= xy * nudge # xy = aoi.centroid ax.annotate(area_of_interest, xy=xy, xytext=xy_nudged, arrowprops=dict(arrowstyle="->"), ha='center') cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {region_of_interest}", fontsize=14); ``` ```{python} area_of_interest = areas_of_interest.iloc[1,0] region_of_interest = areas_of_interest.iloc[1,1] ``` ### `{python} region_of_interest` This map shows Economic Connectedness by **Local Authority District** for the **`{python} region_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') meta_subset = meta_lad.loc[meta_lad['rgn21nm'].str.contains(region_of_interest, na=False)] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.1, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) aoi = meta_subset.loc[meta_subset['lad21nm']==area_of_interest] xy = np.array([aoi.centroid.x.values[0], aoi.centroid.y.values[0]]) nudge = [1, .97] xy_nudged= xy * nudge # xy = aoi.centroid ax.annotate(area_of_interest, xy=xy, xytext=xy_nudged, arrowprops=dict(arrowstyle="->"), ha='center') cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {region_of_interest}", fontsize=14); ``` ```{python} area_of_interest = areas_of_interest.iloc[2,0] region_of_interest = areas_of_interest.iloc[2,1] ``` ### `{python} region_of_interest` This map shows Economic Connectedness by **Local Authority District** for the **`{python} region_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') meta_subset = meta_lad.loc[meta_lad['rgn21nm'].str.contains(region_of_interest, na=False)] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.04, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) aoi = meta_subset.loc[meta_subset['lad21nm']==area_of_interest] xy = np.array([aoi.centroid.x.values[0], aoi.centroid.y.values[0]]) nudge = [1, 1.05] xy_nudged= xy * nudge # xy = aoi.centroid ax.annotate(area_of_interest, xy=xy, xytext=xy_nudged, arrowprops=dict(arrowstyle="->"), ha='center') cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {region_of_interest}", fontsize=14); ``` :::   ## Social Capital is Hyper-local {.smaller} Zooming in even further highlights that social capital can be hyperlocal. Below, we visualise small areas within local authorities to show how social connections vary from neighbourhood to neighbourhood. ::: {.panel-tabset} ```{python} area_of_interest = areas_of_interest.iloc[0,0] region_of_interest = areas_of_interest.iloc[0,1] ``` ### `{python} area_of_interest` This map shows Economic Connectedness by **Middle layer Super Output Area** for **`{python} area_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') # meta_subset = meta_msoa.loc[meta_msoa['lad21nm'].str.contains(area_of_interest)] meta_subset = meta_msoa.loc[meta_msoa['lad21nm'] == area_of_interest] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.2, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {area_of_interest}", fontsize=14); ``` ```{python} area_of_interest = areas_of_interest.iloc[1,0] region_of_interest = areas_of_interest.iloc[1,1] ``` ### `{python} area_of_interest` This map shows Economic Connectedness by **Middle layer Super Output Area** for **`{python} area_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') # meta_subset = meta_msoa.loc[meta_msoa['lad21nm'].str.contains(area_of_interest)] meta_subset = meta_msoa.loc[meta_msoa['lad21nm'] == area_of_interest] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.2, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {area_of_interest}", fontsize=14); ``` ```{python} area_of_interest = areas_of_interest.iloc[2,0] region_of_interest = areas_of_interest.iloc[2,1] ``` ### `{python} area_of_interest` This map shows Economic Connectedness by **Middle layer Super Output Area** for **`{python} area_of_interest`**. Purple values indicate lower Economic Connectedness. Whiter values indicate higher Economic Connectedness. ```{python} variable = 'ec_decile' variable_clean = variable.title().replace('_',' ') # meta_subset = meta_msoa.loc[meta_msoa['lad21nm'].str.contains(area_of_interest)] meta_subset = meta_msoa.loc[meta_msoa['lad21nm'] == area_of_interest] fig, ax = plt.subplots(1,1, figsize = [7,7]) meta_subset.plot( ax = ax, column = variable, cmap=ice_cmap, alpha=0.7, legend=False, edgecolor='black', linewidth=0.2, missing_kwds={ "color": "lightgrey", # fill color for NA "edgecolor": "black", # optional outline for NA polygons "hatch": "///", # optional hatch for NA areas "label": "Missing" # adds entry to legend if legend=True }) # Scale the legend bar divider = make_axes_locatable(ax) cax = divider.append_axes("right", size="5%", pad=0.1) # Create discrete colormap colors = [ice_cmap(i/9) for i in range(10)] # sample 10 discrete colors - using i/9 ensures last colour is used cmap_discrete = mpl.colors.ListedColormap(colors) # Use a simple normalize from 1 to 10 norm = mpl.colors.Normalize(vmin=1, vmax=11) sm = mpl.cm.ScalarMappable(cmap=cmap_discrete, norm=norm) sm._A = [] # Create colorbar cbar = fig.colorbar(sm, cax=cax) # Set custom ticks at bin centers tick_positions = [i + 0.5 for i in range(1, 11)] # 1.5, 2.5, ..., 9.5 cbar.set_ticks(tick_positions) cbar.set_ticklabels(range(1, 11)) # label them 1–10 cbar.set_label('Economic Connectedness: Decile', fontsize=12) cx.add_basemap(ax, crs=meta_subset.crs, source=cx.providers.CartoDB.Positron) ax.set_axis_off(); # ax.set_title(f"{variable_clean}: {area_of_interest}", fontsize=14); ``` :::  ## The Work continues We aim to further explore how these novel data on social capital can be put to use to help improve social outcomes. Stay tuned by subscribing to our newsletter below to hear about our progress.

Exploring Social Capital

The Data

The National View

The Regional Picture

Social Capital is Hyper-local

The Work continues

Footnotes

Stay Up to Date with Our Work