Skip to content

Commit

Permalink
Merge 1e30b17 into 8bf2ff9
Browse files Browse the repository at this point in the history
  • Loading branch information
@edwardoughton
edwardoughton committed Sep 16, 2020
2 parents 8bf2ff9 + 1e30b17 commit 327ebb2
Show file tree
Hide file tree
Showing 19 changed files with 1,732 additions and 683 deletions.
2 changes: 2 additions & 0 deletions .gitignore
View file
Expand Up @@ -112,5 +112,7 @@ vis/backhaul/figures/*
vis/clustering/figures/*
vis/subscriptions/data_inputs/*
vis/smartphones/figures/*
vis/smartphones/data_inputs/*
vis/global_cost/figures/*
vis/tables/*
vis/vis_results/*
232 changes: 182 additions & 50 deletions scripts/preprocess.py
View file
Expand Up @@ -1985,6 +1985,12 @@ def load_subscription_data(path, iso3):

def forecast_subscriptions(country):
"""
Forecast subscriptions.
Parameters
----------
country : string
ISO3 digital country code.
"""
iso3 = country['iso3']
Expand All @@ -1996,7 +2002,7 @@ def forecast_subscriptions(country):
end_point = 2030
horizon = 4

forecast = forecast_linear(
forecast = forecast_subscriptions_linear(
country,
historical_data,
start_point,
Expand All @@ -2019,7 +2025,7 @@ def forecast_subscriptions(country):
return print('Completed subscription forecast')


def forecast_linear(country, historical_data, start_point, end_point, horizon):
def forecast_subscriptions_linear(country, historical_data, start_point, end_point, horizon):
"""
Forcasts subscription adoption rate.
Expand Down Expand Up @@ -2059,82 +2065,208 @@ def forecast_linear(country, historical_data, start_point, end_point, horizon):
return output


def forecast_smartphones(country):
"""
Forecast smartphone adoption.
Parameters
----------
historical_data : list of dicts
Past penetration data.
"""
iso3 = country['iso3']

path = os.path.join(DATA_RAW, 'wb_smartphone_survey', 'wb_smartphone_survey.csv')
survey_data = load_smartphone_data(path, country)

start_point = 2020
end_point = 2030

forecast = forecast_smartphones_linear(
survey_data,
country,
start_point,
end_point
)

forecast_df = pd.DataFrame(forecast)

path = os.path.join(DATA_INTERMEDIATE, iso3, 'smartphones')

if not os.path.exists(path):
os.mkdir(path)

forecast_df.to_csv(os.path.join(path, 'smartphone_forecast.csv'), index=False)

path = os.path.join(BASE_PATH, '..', 'vis', 'smartphones', 'data_inputs')
if not os.path.exists(path):
os.mkdir(path)
forecast_df.to_csv(os.path.join(path, '{}.csv'.format(iso3)), index=False)

return print('Completed subscription forecast')


def load_smartphone_data(path, country):
"""
Load smartphone adoption survey data.
Parameters
----------
path : string
Location of data as .csv.
country : string
ISO3 digital country code.
"""
survey_data = pd.read_csv(path)

survey_data = survey_data.to_dict('records')

countries_with_data = [i['iso3'] for i in survey_data]

output = []

if country['iso3'] in countries_with_data:
for item in survey_data:
if item['iso3'] == country['iso3']:
output.append({
'country': item['iso3'],
'cluster': item['cluster'],
'settlement_type': item['Settlement'],
'smartphone_penetration': item['Smartphone']
})

else:
for item in survey_data:
if item['cluster'] == country['cluster']:
output.append({
'country': country['iso3'],
'cluster': item['cluster'],
'settlement_type': item['Settlement'],
'smartphone_penetration': item['Smartphone']
})

return output


def forecast_smartphones_linear(data, country, start_point, end_point):
"""
Forecast smartphone adoption.
"""
output = []

smartphone_growth = country['smartphone_growth']

for item in data:

for year in range(start_point, end_point + 1):

if year == start_point:

penetration = item['smartphone_penetration']

else:
penetration = penetration * (1 + (smartphone_growth/100))

output.append({
'country': item['country'],
'settlement_type': item['settlement_type'].lower(),
'year': year,
'penetration': round(penetration, 2),
})

return output


if __name__ == '__main__':

# countries = find_country_list(['Africa'])

countries = [
# {'iso3': 'MWI', 'iso2': 'MW', 'regional_level': 2, #'regional_nodes_level': 3,
# 'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 3.5,
# },
# {'iso3': 'UGA', 'iso2': 'UG', 'regional_level': 2, 'regional_nodes_level': 2,
# 'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 2,
# },
{'iso3': 'MWI', 'iso2': 'MW', 'regional_level': 2, #'regional_nodes_level': 3,
'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 3.5, 'smartphone_growth': 1.5, 'cluster': 'C1'
},
{'iso3': 'UGA', 'iso2': 'UG', 'regional_level': 2, 'regional_nodes_level': 2,
'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 2, 'smartphone_growth': 1.5, 'cluster': 'C1'
},
{'iso3': 'SEN', 'iso2': 'SN', 'regional_level': 2, 'regional_nodes_level': 2,
'region': 'SSA', 'pop_density_km2': 200, 'settlement_size': 1000, 'subs_growth': 1.5,
'region': 'SSA', 'pop_density_km2': 200, 'settlement_size': 1000,
'subs_growth': 1.5, 'smartphone_growth': 1.5, 'cluster': 'C2'
},
{'iso3': 'KEN', 'iso2': 'KE', 'regional_level': 2, 'regional_nodes_level': 1,
'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 1.5,
'region': 'SSA', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 1.5, 'smartphone_growth': 1.5, 'cluster': 'C2'
},
{'iso3': 'PAK', 'iso2': 'PK', 'regional_level': 3, 'regional_nodes_level': 2,
'region': 'S&SE Asia', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 2.5, 'smartphone_growth': 2, 'cluster': 'C3'
},
# {'iso3': 'PAK', 'iso2': 'PK', 'regional_level': 3, 'regional_nodes_level': 2,
# 'region': 'S&SE Asia', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 2.5,
# },
{'iso3': 'ALB', 'iso2': 'AL', 'regional_level': 2, 'regional_nodes_level': 2,
'region': 'Europe', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 0.3,
'region': 'Europe', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 0.3, 'smartphone_growth': 2, 'cluster': 'C4'
},
{'iso3': 'PER', 'iso2': 'PE', 'regional_level': 2, 'regional_nodes_level': 1,
'region': 'LAC', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 0.5, 'smartphone_growth': 2, 'cluster': 'C5'
},
{'iso3': 'MEX', 'iso2': 'MX', 'regional_level': 2, 'regional_nodes_level': 1,
'region': 'LAC', 'pop_density_km2': 500, 'settlement_size': 1000,
'subs_growth': 1.2, 'smartphone_growth': 2, 'cluster': 'C6'
},
# {'iso3': 'PER', 'iso2': 'PE', 'regional_level': 2, 'regional_nodes_level': 1,
# 'region': 'LAC', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 0.5,
# },
# {'iso3': 'MEX', 'iso2': 'MX', 'regional_level': 2, 'regional_nodes_level': 1,
# 'region': 'LAC', 'pop_density_km2': 500, 'settlement_size': 1000, 'subs_growth': 1.2,
# },
]

for country in countries:

print('Working on {}'.format(country['iso3']))

print('Processing country boundary')
process_country_shapes(country)
# print('Processing country boundary')
# process_country_shapes(country)

# print('Processing regions')
# process_regions(country)

print('Processing regions')
process_regions(country)
# print('Processing settlement layer')
# process_settlement_layer(country)

print('Processing settlement layer')
process_settlement_layer(country)
# print('Processing night lights')
# process_night_lights(country)

print('Processing night lights')
process_night_lights(country)
# print('Processing coverage shapes')
# process_coverage_shapes(country)

print('Processing coverage shapes')
process_coverage_shapes(country)
# print('Getting regional data')
# get_regional_data(country)

print('Getting regional data')
get_regional_data(country)
# print('Generating agglomeration lookup table')
# generate_agglomeration_lut(country)

print('Generating agglomeration lookup table')
generate_agglomeration_lut(country)
# print('Load existing fiber infrastructure')
# process_existing_fiber(country)

print('Load existing fiber infrastructure')
process_existing_fiber(country)
# print('Estimate existing nodes')
# find_nodes_on_existing_infrastructure(country)

print('Estimate existing nodes')
find_nodes_on_existing_infrastructure(country)
# print('Find regional nodes')
# find_regional_nodes(country)

print('Find regional nodes')
find_regional_nodes(country)
# print('Fit edges')
# prepare_edge_fitting(country)

print('Fit edges')
prepare_edge_fitting(country)
# print('Fit regional edges')
# fit_regional_edges(country)

print('Fit regional edges')
fit_regional_edges(country)
# print('Create core lookup table')
# generate_core_lut(country)

print('Create core lookup table')
generate_core_lut(country)
# print('Create backhaul lookup table')
# generate_backhaul_lut(country)

print('Create backhaul lookup table')
generate_backhaul_lut(country)
# print('Create subscription forcast')
# forecast_subscriptions(country)

print('Create subscription forcast')
forecast_subscriptions(country)
print('Forecasting smartphones')
forecast_smartphones(country)

0 comments on commit 327ebb2

Please sign in to comment.