1. No More “Half-Fast”: Improving
US Broadband Download Speed
Georgetown University – 2015 Data Science Capstone
Brittne Nelson PhD, Amgad Sirag, Ernest S.
2. Approach and Overview
What?
• Broadband Data Story
• Research Problem
• Data Science Pipeline
So What?
• Data Visualization Story
• Findings
• Lessons Learned
Now What?
• Future Research
• Conclusions
4. There were
communities
with no
broadband
access
Every day,
residents and
businesses had
limited or no
access to
resources,
services,
content, new
customers, and
new technology
limiting
opportunities and
community
empowerment
One day, the
US government
created the SBI
to facilitate the
integration of
broadband and
information
technology into
state and local
economies
Because of
that, states
did more to
quickly
expand
broadband to
more areas
Because of
that, the SBI,
decision
makers, and
researchers-
including us-
were able to
assess how
broadband is
being
implemented
across the
US
Until finally,
residents and
businesses
gained more
access to
resources
services,
content, new
customers,
and
technology
that
empowered
and gave
them a
competitive
edge
Data Story
5. Benefits of Broadband
• Increased job opportunities
• Increased employment opportunities due to telework
• Higher pay
• Increased economic security
• Recruitment of job seekers, especially in rural areas
• Increased access to and quality of healthcare
• Availability of a wide variety of entertainment
• Increased participation in everyday economic, social, and community life
• Improved social connections to existing friends and acquaintances
• Creation of new relationships based on common interests
• Improved social integration of minority populations
• More positive attitudes toward aging
• Higher levels of perceived social support and connectivity among seniors
• Lower prices for online purchases
• Improved variety of items available for purchase
• Better purchasing decisions based on online information
• Savings in time and money for online vs. paper-based activities
• Improved connectivity for social or political action
Sources: Center for Social Inclusion,. (2010). The Promise and Challenge of Community Broadband Models. New York City: Center for
Social Inclusion.
Analytics ASR,. (2014). Final Report: Social and Economic Impacts of the Broadband Technology Opportunities Program. Potomac
Maryland.
6. Research Problem
• Does
broadband
availability
and
speed
make
a
state’s
economy
and
it’s
residents
compe55ve?
• When
will
every
state
reach
98%
broadband
connec5vity?
• How
are
community
economic
features
impac5ng
or
related
to
broadband
development?
7. Hypotheses
• Broadband
speed
and
accessibility
will
cluster
in
urban
areas
• Areas
with
more
broadband
speed
will
have
lower
unemployment,
more
businesses,
and
larger
popula5ons
• Broadband
growth
is
not
consistent
across
all
coun5es
• Based
on
past
growth,
broadband
coverage
is
not
expected
to
be
available
in
98%
of
all
coun5es
in
2016
8. Data Sources
• Na5onal
Broadband
Map
Maximum
and
Minimum
Download
Speed
by
County,
June
2011-‐June
2014
– Na5onal
Telecommunica5ons
and
Informa5on
Administra5on
hRp://www.broadbandmap.gov/data-‐download
• Labor
Force
Data
by
County
Annual
Average,
2011-‐2013
– U.S.
Department
of
Labor
Local
Area
Unemployment
Sta5s5cs
hRp://www.bls.gov/lau/
• Demographic
Popula5on
by
County,
2010
– U.S
Census
Bureau
hRp://fac[inder.census.gov/faces/nav/jsf/pages/index.xhtml
• Total
Number
of
Business
Establishments,
2011-‐2012
– U.S
Census
Bureau
hRp://fac[inder.census.gov/faces/nav/jsf/pages/index.xhtml
13. Hypotheses
Results
• Broadband
speed
and
accessibility
will
cluster
in
urban
areas
• “URBAN”
TOO
DIFFICULT
TO
DEFINE
GIVEN
PROJECT
TIMELINE,
NOT
ANALYZED
• Areas
with
higher
broadband
speed
have
lower
unemployment,
more
businesses,
and
larger
popula5ons
• NOT
TRUE
• Broadband
growth
is
not
consistent
across
all
coun5es
• TRUE
• Based
on
past
growth,
broadband
coverage
is
not
expected
to
be
available
in
98%
of
all
coun5es
in
2016
• NOT
ENOUGH
DATA
TO
COMFORTABLY
FORECAST
14. Summary of Findings
• Iden5fied
economic
features
are
mild
drivers
of
technology
implementa5on
specifically
broadband
speed.
• Broadband
availability
makes
a
state
economy
and
it’s
residents
compe55ve.
• Implemen5ng
broadband
is
not
the
silver
bullet
to
community
development
or
economic
growth,
it
should
be
incorporated
with
other
economic
and
social
features.
15. Lessons Learned
• Quantity of data is important for forecasting
• Source of data is important. SBI reports data from
providers which makes it somewhat difficult to
assess
• Plan a significant amount of time for data
wrangling
• Master each step of the data science pipeline
before moving on
• Operationalize more factors to provide a clear
picture of relationships when identifying
hypotheses
17. Future Research
• Develop a matched pairs analysis framework that
compares changes in the availability of broadband at the
state level between counties
• Measure how much of the growth in availability within
these counties occurred due to funding (Grants, Federal
Government, Private Organizations)
• Examine broadband’s long-term quantitative
extrapolations and impact on social and economics
• Index and model additional community factors such as
education, adoption, tax rate, etc in order to broadly
define economic impact
18. Conclusions
• There is a business case for continued focus on
broadband improvement
• Broadband improves the overall communities
• Drives economic development and shared
opportunities
• Improve quality of life across the United
States
19. Thank You to the Georgetown University 2015 Data Science Program
Faculty
Benjamin Bengfort
Allen Leis
Sacha Litman
Laura Lorenz
Salil Mehta
Tony Ojeda
(and lady!)