28 ISE Magazine | www.iise.org/ISEmagazine
Hunger is a reality for more than 50 million peo-
ple in America, 1 in 6 of the U.S. population,
including more than 1 in 5 children. Though
hunger is a sensation that can be felt physically,
it can be laborious to measure it. One such way
to measure hunger and food scarcity of a large
population is by food insecurity.
Food insecurity is the state or condition where someone
is living short of obtaining ample amounts of cost-effective,
accessible and nourishing food. In 2013, households that
had higher rates of food insecurity than the national av-
erage included households with children (20%), especially
those households with children headed by single women
(34%) or single men (23%).
Also, as per the 2015 statistics provided by Feeding Amer-
ica ( feedingamerica.org), one of the largest domestic hunger-
relief organizations in the United States, 42.2 million
Americans live in food insecure households, including 29.1
million adults and 13.1 million children. This constitutes
about 13.1% of households that are food insecure.
To curb this issue, food insecure individuals and families
receive assistance from the government and from philan-
How data science can provide
nourishment and food
Analytics boosts food banks’ efforts in fundraising, distribution processes
By Rahul Srinivas Sucharitha and Seokcheon Lee
December 2021 | ISE Magazine 29
thropic programs. One such program is Feeding America,
which provides food and assistance to these individuals
through a nationwide network of around 200 food banks
and around 63,000 food pantries and meal programs. Of
those, 66% of food pantries, 41% of kitchen programs and
11% of shelter programs have no paid staff and rely entirely
on volunteers.
Food bank network members of Feeding America supply
food to more than 37 million Americans each year, includ-
ing 14 million children and 3 million seniors; 34% of all
households served by Feeding America have had to choose
between paying for food and paying for medicine or medi-
cal care. Additionally, 72% of food banks do not feel as
though they are able to adequately meet the needs of their
communities without adjusting the amount of food distrib-
To assist in their efforts, industrial engineers can use
their skills to create a realistic food allocation metric and
ensure equitable and nutritious food to those who are
food insecure.
How food banks need data analytics
Food banks obtain donated food and grocery products
from national food and grocery manufacturers and retail-
ers. These donations then are distributed to food pantries
and meal programs based on their availability. People in
need receive the donated foods from these food pantries
and meal programs. Food is usually received from food do-
nors by following a fixed schedule that is convenient to the
Once the food is obtained from the donors, it is processed
and inspected at the food bank before being made available
for distribution to the agencies (Figure 1). The food is dis-
tributed to the agencies by trucks that are maintained by
the banks or are being rented based on the travel schedule.
Agencies receive the donated food based on their storage
capacity (in pounds). It is then prioritized based on distance
from the food bank; usually the nearest agencies are given
priority. This leads to circumstances where an inadequate
amount of nutritious food is provided to people who need
it as the food is distributed based on weight and not based
on the amount of the type of food, such as protein or fats.
According to the recent Salesforce Nonprofit Trends Report,
more than half of the nonprofits (53%), including food as-
sistance programs and other types of nonprofits, find it easy
to collect program data. However, putting that data into
action is more complicated. Fewer than half (47%) say it is
easy to analyze that data, leading to a wide range of strug-
gles when it comes to tracking and quantifying things like
Food bank distribution
How donated food is passed from donors to food banks and then on to participating agencies.
42.2 million Americans live in food
insecure households, including 29.1
million adults and 13.1 million
children. This constitutes about 13.1%
of households that are food insecure.
30 ISE Magazine | www.iise.org/ISEmagazine
How data science can provide nourishment and food security
impact and performance. Also, just 41% say they find it easy
to have data power the overall impact of programs.
In addition, despite the fact that mobility is becoming
more important to nonprofits each passing year, only 29%
say they can easily capture and access data from a mobile
device. This shows the importance of analytics tools for
nonprofits, but only 45% currently use analytics, according
to the report, with an additional 30% planning to use such
tools within two years.
Nonprofits have huge datasets they can exploit to de-
How food banks began
The food bank concept was started by John van Hengel in Phoenix in the late 1960s (Figure 2). The retired businessman was volunteering
at a soup kitchen trying to find food to serve the hungry and observed the discarding of edible food from grocery stores. He concluded
the food could be used to feed those in need by storing the items for people to pick up.
Van Hengel established St. Mary’s Food Bank in Phoenix as the nation’s first food bank. In its first year, he and his team of volunteers
distributed 275,000 pounds of food. Word of their success spread quickly and others in the country began to take note. In 1976, the U.S.
government provided van Hengel’s food bank with a grant to assist in developing similar operations around the country. That was helped
by the passage of the 1976 Tax Reform Act that made it financially advantageous for companies to donate their products. This federally
funded development lengthened and ultimately incorporated to manage the solicitation of assistance from donors and develop standards
for food banks pertaining to storage capacity, quality control and management.
By 1977, food banks were established in 18 cities across the country. As the number increased, van Hengel created a national
organization for them, and in 1979 established Second Harvest, later known as America’s Second Harvest the Nation’s Food Bank
Network. Federal funding was discontinued by 1982, so America’s Second Harvest pursued alternative financial sources and moved its
national office to Chicago in 1984. The organization continued to grow as the practice of food banking gained acceptance along with
support from the food industry and local social service providers.
With many major cities establishing food banks by the mid-1980s, network expansion slowed and America’s Second Harvest’s shifted
its focus to improving existing programs. The professionalism and efficiency of food bank operations improved dramatically, resulting in
a much greater amount of food and products being distributed by the network.
In 1999, the name of the national organization was officially changed to America’s Second Harvest with a goal of ending hunger in
America. In March 2000, it merged with Food Chain, the nation’s largest food rescue organization. That produced the country’s most
comprehensive and efficient charitable food rescue and distribution organization.
In 2008, the network changed its name to Feeding America to better reflect the mission of the organization. Today, it is the nation’s
largest domestic hunger-relief organization with an efficient network of 200 food banks nationwide. As food insecurity rates hold steady at
the highest levels ever, the Feeding America network of food banks has risen to meet the need serving the entire U.S., with its food banks
operating 63,000 agencies that address hunger through emergency food assistance and programs. Learn more at feedingamerica.org.
Food bank history
A timeline details the origin of food banks.
December 2021 | ISE Magazine 31
velop statistical models that can help them op-
timize fundraising. Using segmentation and
predictive modeling helps them identify and
target the right groups to develop a more re-
ned marketing strategy. Data insights help
nongovernmental organizations identify and
segregate donors based on various factors,
which they can use to drive their marketing
and fundraising efforts more effectively. Data
analysis also helps charitable organizations dis-
cover relationships that can help them develop
specific incentives.
Nonprofits such as food banks and agencies
need to sustain themselves throughout the en-
tire year. To do so, they need to be efficient
in their operations and donor outreach efforts.
Data analytics tools can help on those fronts,
and recent research indicates that such solutions
are needed at nonprofits. Food assistance programs can use
analytics to streamline operations, increase cost efficien-
cy, determine and optimize financial margin by program,
model and forecast performance (e.g., membership trends,
donor trends, resource needs and revenue expectations),
improve the budgeting process, and enhance overall mis-
sion effectiveness. Such solutions can provide significant
benefits to nonprofits, including the ability to measure a
return on investment.
Analytics can also help nonprofits figure out the best
prospects to engage to make a donation. Modeling can be
used to examine donor demographics, giving data and in-
teractions with an organization to predict the future giving
behaviors of potential donors. Wealth screening and pub-
licly available information can then be used to determine
the prospects with the most money to potentially seek for
Used together, these tools can help an organization nar-
row the scope of the prospect database and identify pros-
pects for assignment to the development staff, make more
informed donation requests and identify new prospects to
ll the donor pipeline. Analytics can also help strengthen
an organizations membership recruitment and fulllment
operations. Data analysis and visualization can also play a
huge role in real time tracking during crises by optimizing
relief efforts.
Traditionally, food assistance programs have purchased
lists of names in the target market and sent communications
deep into that list, as well as to their existing membership
base. However, instead of using that blunt approach to yield
new members, nonprofits can use data analytics to analyze
various factors, activities and characteristics in concert to
determine how they relate to and affect each other. This
enhanced approach enables organizations to generate an
optimal number of engaged members in a fiscally prudent
manner. Using data analytics, planning exercises can con-
sider a more complete set of factors, such as membership
outreach yield, cost per outreach, market penetration, de-
mographic factors and changes, membership engagement
and membership product cross-selling.
There are numerous other benefits to analytics. The tools
can be used to track and analyze nonprofit staff activity and
motivate them through data-driven performance bench-
marks. Basic activity metrics, including donor visits, solici-
tations, gifts received and new prospects, will provide the
baseline health of the donor pipeline and enable revenue
forecasting. Analytics can also help improve donor relations
by providing more quantiable data to donors on the im-
pact of their “investment,” which may then motivate them
to give again. Organizations must not only make data mea-
surement a priority but also understand the various ways
Nonprofits such as food banks and
agencies need to sustain themselves
throughout the entire year. To do
so, they need to be efficient in their
operations and donor outreach efforts.
Data analytics tools can help
on those fronts, and recent research
indicates that such solutions are
needed at nonprofits.
32 ISE Magazine | www.iise.org/ISEmagazine
How data science can provide nourishment and food security
that impact data can be effectively communicated.
Data analysis can help these organizations make sure that
theyre putting their time, effort, money and energy into
the right channels. Nonprofit organizations need funds to
fulll their missions, but they also need to prove the results
of their work to attract donors. Data science has immense
potential in optimizing this cycle and can help organiza-
tions make well-informed, sound decisions. While not ev-
ery organization can afford to hire full-time data analysts,
some fill the resource gap by seeking help on platforms like
Kaggle and Stack Exchange, or by hiring freelance data
scientists and consultants through Kolabtree and Experfy.
That data science bears enormous potential for govern-
ment organizations and nonprofits, big and small, is quite
clear. But accessing experts is a common problem that these
organizations face. Processes often are bureaucratic, and the
funds collected may be inadequate to employ a full team.
Thus, many organizations dont see the power of data sci-
ence manifest.
One solution for organizations is to consult freelance data
science experts for short-term projects. This helps them ac-
cess the expertise of data analysts while keeping the process
cost-effective. This is one practical way the third sector can
reap the benefits of data science and stay empowered, so
that they can empower others.
Filling the need for nutrition
Regional food banks collect the donated, purchased and
government-supplied foods, and distribute the foods to
their 63,000 locally afliated agencies that provide gro-
ceries and hot meals to low-income families through meal
programs such as food pantries and soup kitchens. The core
principles governing the distribution of foods to the agen-
cies are “as much as possible” and “as equitable as possible.
Accordingly, the distribution performance of a food
bank is typically evaluated in total pounds distributed and
distribution equity among counties (or agencies). We pose
two fundamental issues regarding these distribution per-
formance metrics that naturally drive the way food distri-
bution operations are preconditioned and executed across
strategic, tactical and operational levels.
First, due to the lack of appropriate trade-off between
the two metrics of total and equity, various distribution
strategies can be, and have been, applied, occasionally pro-
ducing irrational and/or confusing outcomes. For instance,
consider three agencies responsible for equal-sized food-
December 2021 | ISE Magazine 33
insecure populations. Taking more into account that the
equity in food distribution can lead to preferring value
amounts in pounds of food of [100 100 100] to [100 150
100], though the latter clearly dominates the former. This
problem stems from the separate pursuit of these possibly
conicting metrics; thus, an integrated approach is required
to eliminate such irrational preferences.
Second, as mentioned earlier, the distribution perfor-
mance is based on the quantity (weight) only despite the
fact that the nutritional quality of the foods distributed by
food banks can make a critical difference to the health of
recipients. This further leads to an unreasonable outcome
from the aforementioned instance. Together with concerns
about obesity and diet-related chronic diseases such as dia-
betes, food banks are increasingly aware of the need to im-
prove the nutritional quality of charitable foods in the form
of nutrition proling to quantitatively score the nutritional
value; nutrition policies to guide efforts to eliminate un-
healthy products such as soda or candy; or increasing the
amount of fresh produce to fill the nutritional gaps.
However, no attention has been paid to date on how to
distribute foods in a nutrition-rich or nutrition-balanced
manner. As shown in an observational study with 269 food
pantries supplied from two large food banks in Minnesota
in 2013, the nutrition quality (measured in Healthy Eating
Index, HEI-2010) ranged from 28 to 82 out of 100 (Ap-
plication of the Healthy Eating Index-2010 to the Hunger
Relief System,” Marilyn S. Nanney, Katherine Y. Gran-
non, Colin Cureton, Courtney Hoolihan, Mark Janowiec,
Qi Wang, Cael Warren and Robert P. King 2, 2016). This
large variability could be because of the lack of systematic
consideration of nutrition quality in the current metrics of
distribution performance.
Where data scientists fit in
If data scientists begin to invest their efforts in areas beyond
business industries, they will discover how far and wide
their skills and experience can be helpful. Nonprofit orga-
nizations are mostly run with the help of funds. It is incred-
ibly difficult to gain these funds and success depends on
factors like location and the cause the nonprofit supports.
One of the most important accomplishments nonprofits
must achieve is to prove their efforts for a certain cause are
impactful, resulting in positive change. This attracts more
people to provide funds and voluntary efforts. If nonprofit
organizations could get access to the wonders of data sci-
ence, the process of getting funds and proving the impact
could be optimized. Better optimization will obviously
produce a more positive result for all involved. Hence,
many other nonprofits apart from food assistance programs
are applying data science and machine learning techniques.
With the understanding that everything that is or can
be digitized is data, organizations can start to mine data to
learn about their audiences, their approaches and the con-
text of the world around them. The fact that we still have
long-standing problems – societal challenges that nonprofit
organizations are set up to address – means we still have
more work to do, and big data might well be a lever to
creating change.
Admittedly, a data science team is not easy to afford.
Moreover, most of the talent is centered in the business
sector. It is time we admit that this gap of resource needs to
be overcome. In the same line of thought, it should be men-
tioned that big data and data science can provide not only
smarter solutions at the consumer end but actually bring a
positive change for humanity. Hence, several of these orga-
nizations can benefit from ISEs providing pro bono train-
ing, enablement and teaching.
If you are a data scientist, try locating volunteer projects
near you. You will find this rewarding, and together we can
help in providing nourishment and relief today and work
toward providing food security tomorrow, overcoming
hunger and malnutrition.
Additionally, industrial engineers can aid in ensuring a
realistic food allocation metric for the people in need. By
designing different metrics that aggregate quantities and
qualities of distributed foods and developing techniques to
evaluate and compare the potential metrics in simplistic and
complex scenarios where a mix of foods needs to be allo-
cated to a set of entities, we can provide a suitable metric to
optimally allocate the foods. The allocation results will go
through analysis to identify the strengths and weaknesses of
each of the metrics.
Rahul Srinivas Sucharitha is a Ph.D. candidate in the School of
Industrial Engineering at Purdue University in West Lafayette,
Indiana. He earned a bachelors degree in mechanical engineer-
ing from the College of Engineering, Guindy, Chennai, India, in
2013 and a masters degree in industrial engineering from Purdue
University in 2015. His research interests include modelling com-
plex systems, food bank distribution networks and logistics.
Seokcheon Lee is an associate professor in the School of Industrial
Engineering at Purdue University in West Lafayette, Indiana.
Lee earned bachelor’s and master’s degrees in industrial engineer-
ing from Seoul National University, Seoul, Korea, in 1991 and
1993, respectively, and a Ph.D. in industrial engineering from
Pennsylvania State University in 2005. His research work has
been focused on distributed control of large-scale complex networks
– supply network coordination, dynamic resource allocation and
emergency logistics – based on the fundamental decision principles
designed from multidisciplinary perspectives. He has been an IISE
member since 2009.