Using a Mobile Based Web Service to Search for Missing People – A Case Study Of Kenya

International Journal of Computer Applications Technology and Research
Volume 4– Issue 7, 507 - 511, 2015, ISSN:- 2319–8656
www.ijcat.com 507
Using a Mobile Based Web Service to Search for Missing
People – A Case Study Of Kenya
Thomas M. Omweri
School of Computing and Informatics,
University of Nairobi, Kenya
Andrew M. Kahonge
School of Computing and Informatics,
University of Nairobi, Kenya
Abstract: Being out of touch with a loved one is concerning and not hearing from someone you care about is terrifying. Several cases
of missing people have been reported for many years, where most of the searches turn out unsuccessful. In order to quickly reunite
families and friends with their missing loved ones, a solution for effectively searching for the missing people is presented. In
evaluation of this solution, an F1 score test was simulated using 20 scenarios, out of which an impressive score of 0.72 was attained.
The study concludes that we need to leverage on mobile based technology to device a more efficient method of finding missing
persons more easily and quickly.
Keywords: Mobile Application; Emergency communication system; National Disaster Operation Centre(NDOC); Emergency
communication system(ECS); United Nations(UN); Missing Persons Community of Interest(MPCI); International Committee of the
Red Cross(ICRC)
1. INTRODUCTION
Reports of missing persons worldwide have increased
significantly in the past recent years, from roughly 450,000 in
1990 to about 10,000,000 this year [1]. The increase was
driven in part by the ever growing population. The numbers
indicate that more people are becoming victims each day. An
astounding 2,300 Americans are reported missing every day,
including both adults and children. Kenya on the other hand
has at least 20,000 missing people on record every year. Out
of the reported number,40% are located after a long period of
search while 30% are left untraced. Only 30% of the reported
victims are found within a reasonably short period of up to 3
months.
More recently, the abductions of children and adults have
reawakened public concern about missing people. In most
parts of the world, the police and non-governmental
organizations working with missing people have recently
reviewed their policies and are planning to improve
coordination of their work [2]. People end up missing in
different scenarios [3]. The circumstances that may lead
adults or children to become missing people are often
complex and multi-layered. The missing phenomenon is best
understood as a continuum in which a break in contact may be
either intentional or unintentional. Some people make a
conscious decision to leave, albeit often not in circumstances
of their own choosing, while others may drift apart from
family members over time. Some may never have intended to
be missing, and indeed may not conceptualise their experience
in these terms, while others may be forced apart through the
actions of others. Some of the causes entailed herein are
natural disasters, psychological complications, abduction and
domestic conflicts [4].
2. PREVIOUS EFFORTS DONE IN
THE SEARCH FOR MISSING PEOPLE
Research concerning missing persons has been done in the
past. A few of the research efforts have been successfully
implemented while others did not see the light of the day for a
number of reasons. These past researches can enable us to
develop a lens through which we can view the phenomenon
under this important study [5].
Advances in technology have had a major impact on tracing,
mainly by speeding up the transmission of information to
huge numbers of people, according to the International
Committee of the Red Cross (ICRC) Central Tracing Agency.
The ICRC started tracing in the late 1800s to alert families to
the whereabouts and well-being of detained relatives. It
currently relays hundreds of thousands of messages linking
families back together and providing the peace of mind and
closure so often absent in times of crises. In 2009 alone, more
than 253,000 messages were collected and delivered. Tracing
assisted the repatriation of Congolese prisoners of war, and
enabled nearly 200 video calls between detainees and their
families[6].
Following Haiti’s earthquake in January 2010, Google
developed an open source web application, Person Finder,
which is a registry and message board for survivors, family
and friends to post and search for information about one
another's whereabouts following a natural disaster. Up until
now, following five natural disasters, the registry has
collected more than 200,000 victim names [7].
The Dutch government has also adopted a mobile phone
danger alert system that sends text messages to people who
could be affected by natural disasters or terrorist attacks. The
system, called Cell Broadcast, uses GSM technology to
identify cell phone users in a particular area [8]. If a disaster
occurs, a message is sent to all phones in the area, warning of
the danger.
3. METHODOLOGY
The goal was to come up with a prototype of a solution for
finding missing persons fast enough to find them safe and
sound. The solution is an innovation leveraging on the use of
the readily available mobile phone devices and the internet.
This kind of approach has shown success in the past [9]. It is
also intended to be a solution that takes into consideration
privacy and other legally constraining issues that surround
missing people [10]. The block diagram in figure 1 below
represents the conceptual model of the solution.

www.ijcat.com 508
Figure 1. The Conceptual model
2.1 Reporting a Missing Person
In case a person goes missing, family members, friends or
acquaintances should be able to report the case on the system.
They should be able to do so by registering the person's
details such as name, age, tribe, place of origin and
description. This should give an exhaustive description of the
missing person to increase chances of the reader spotting and
reporting them.
2.2 Reporting a Found Person
A person having been reported as missing, can be reported as
found if spotted anywhere. Anyone with the leading
information should be able to post them on the system, about
the person in question.
Even the missing people can report themselves in case they
are able to access the system and in a condition that may
allow them to do so. The current location and contact details
of the missing person should be provided, as well as those of
the person reporting the case. The reporter may need to be
contacted for further details.
2.3 Sending Email Notification Alerts
In case there is a match between a missing and a found person
the system will send email alerts to those who have reported
missing cases and have subscribed to the service. The
algorithm used to match involves some major details like
names, gender and tribe of the missing person.
2.4 Search for a Missing Person
Users should be able to search for their missing loved ones on
the system. The system provides a search criteria to make
their search easier and relevant e.g name of missing person,
tribe, age and gender.
The user is then able to see feeds or updates about the missing
person they reported. There may be multiple entries about a
single missing person, reported by different people at different
times. All these updates should appear if they are associated
with the missing person, as they may provide quality leading
information to finding the missing person. Figure 2 below is a
screen shot that captures the search results.
Figure 2. Screen shot showing search results of missing
people
3. GENERAL PROCEDURE
The general procedure followed by the system is as illustrated
in figure 3 below. A family member or friend reports a
missing person case. Anyone with recovery details about the
missing person updates the records by providing leading
information. The missing case reporter keeps searching for
any leading information from the system.

Volume 4–
Fig 3. General procedure flow chart
The stakeholders of this emergency communication system
include the lost case reporter, found case reporter and missing
person. The lost case reporter may be a friend, family member
of a person acquainted to the missing person. The interactions
of the stakeholders with the system is depicted in the use case
diagram in figure 4 below.
Fig 3. General procedure flow chart
of a person acquainted to the missing person.
4. RESULTS
After full implementation and testing of the system,
evaluation of the prototype was done with the aim to
determine if the developed system is delivering the expected
results. The following areas were evaluated to provide
answers to the research questions set at the feasibility study of
the project, which are in line with the project objectives and
requirements. This information was realized by use of both
qualitative and quantitative methods during the co
data [11].
4.1 Determining the search success rate
To determine this rate, 20 people were reported as missing in
the prototype. The success rate was recorded in the database
clearly showing matches of people reported as missing and
those reported as found. Whenever there is a match, a
notification alert is sent to the reporter of the missing person
together with the leading information concerning the victim's
current whereabouts. Table 1 below is a summary of the
evaluation results.
– Issue 7, 507 - 511, 2015, ISSN:- 2319–8656
of a person acquainted to the missing person. The interactions
of the stakeholders with the system is depicted in the use case
After full implementation and testing of the system,
evaluation of the prototype was done with the aim to
determine if the developed system is delivering the expected
luated to provide
answers to the research questions set at the feasibility study of
the project, which are in line with the project objectives and
requirements. This information was realized by use of both
qualitative and quantitative methods during the collection of
Determining the search success rate
To determine this rate, 20 people were reported as missing in
the prototype. The success rate was recorded in the database
clearly showing matches of people reported as missing and
ported as found. Whenever there is a match, a
notification alert is sent to the reporter of the missing person
together with the leading information concerning the victim's
current whereabouts. Table 1 below is a summary of the
Table 1. Summary of the evaluation results of finding
missing persons
Cases reported Search success
20 16
The following pie chart in figure 4 illustrates the portion of
success rate versus the failure rate.The success rate is
significant enough to quaify the prototype as having satisfied
the objectives and expectations of the study.
Figure 5. Pie Chart showing the success rate of finding
missing people by using the mobile application
During the tests carried out in evaluation it was determined
that reasons why a missing person may not be identified are:
• The person is not reported as missing in the system.
• The person may be reported with different names
and other details from the ones used to search.
The prototype was tested under each of the fo
scenarios:
• Unit testing - Each functional module was tested
during and after development to ensure that it meets
the requirements. Additionally, basic validation has
been done to ensure the correct input data on each
module.
• Integration testing - This testing
during and after integration of all the modules. It
therefore checks that the system has the flow
required, from the point of reporting lost cases to
missing cases and outputting notification alerts.
Summary of the evaluation results of finding
Search failure
4
The following pie chart in figure 4 illustrates the portion of
success rate versus the failure rate.The success rate is
to quaify the prototype as having satisfied
the objectives and expectations of the study.
Figure 5. Pie Chart showing the success rate of finding
missing people by using the mobile application
During the tests carried out in evaluation it was determined
that reasons why a missing person may not be identified are:
The person is not reported as missing in the system.
The person may be reported with different names
and other details from the ones used to search.
The prototype was tested under each of the following
Each functional module was tested
during and after development to ensure that it meets
Additionally, basic validation has
been done to ensure the correct input data on each
This testing was done before,
during and after integration of all the modules. It
therefore checks that the system has the flow
required, from the point of reporting lost cases to
missing cases and outputting notification alerts.

www.ijcat.com 510
• Acceptance testing - We conducted a beta testing
of the system on a sample of users. Some of the
users disguised as lost case reporters while others as
found case reporters. Wherever there is a match of
records, email notifications are sent to the lost case
reporters, providing them with the leading
information about their missing people. The search
function was also included in the test.
Besides, a computation of F1 score test was conducted to
measure the system's accuracy. This being a statistical
analysis [12], the F1 score considered both precision p and the
recall r of the test to compute the score. Precision(p) is the
number of correct results divided by the number of all
returned results while recall(r) is the number of correct results
divided by the number of correct results that should have been
returned. The F1 score can be interpreted as a weighted
average of the precision and recall, where an F1 score reaches
its best value at 1 and worst score at 0.
After considering 20 search test cases in the system, the
following results were yielded:
Total scenarios = 20
Successful searches = 16
Correct searches = 13
Precision = 13/16
Recall = 13/20
F1 = 2(0.528125/1.4626)
F1 = 0.72
4.2 Measuring access of web database
The measures of efficiency considered were:
• Successful lost case report
• Successful found case report
• Successful search for reported case
• Success in receiving notification alert emails
Twenty entries were done and data on the above measures
was taken and recorded on Mysql server database table. Table
6 below is a summary of the client access efficiency data as
analyzed.
Table 2. Database client access success measure
Access Efficiency Measures Success Failure
Lost case report 20 0
Found case report 20 0
Search for reported person 16 4
Notification alerts 16 4
4.3 Measuring rate of success in the identity
levels of missing people
The users suggested that the system should include photos of
the missing people to be able to identify them more easily.
The use of a photo is significant as some people may
recognize the missing person by view of their photo even
when they do not have the additional descriptive details like
the name.
It was also suggested that the identification details need to be
more flexible and exhaustive in order to increase chances of
identifying a missing person. For instance, the missing
person's age should be a range of numbers rather than an
absolute number because sometimes it's not easy to tell the
exact age of a person. However, an age bracket consisting of a
range of years may work better in this case for instance 25-30
years.
4.4 Discussion of results
The results indicate that the system accuracy is high when
reporting the lost and found cases. This is simply because all
that occurs at this point is to fill in the respective forms and
submitting them.
However when it comes to searching for the missing persons,
the accuracy goes down by 20 % because this step involves a
search algorithm that takes into considerations many
parameters. In the event that some parameters used during the
search do not match with those used during the reporting of
the lost or found case, the intended result returns null even
when it should have retrieved the record. This results into a
false negative. This is a common phenomenon in social
research methods [13].
Consequently, the email alerts or notifications are affected by
the result of the preceding step. The alerts step is equally
affected by 20% and does not send notifications to all the
recipients as it should.
In the general overview, the system achieves an accuracy
level of approximately 80% which is impressive. Even in
cases where a false positive or false negative is returned,
repeated search with different parameters may increase
chances of returning the desired true results.
From the results we can compare the performance of the new
system with the pre-existing systems in the same domain.
Being a mobile based application, its more convenient and
accessible in comparison with the web based solutions like
Google's person finder. The new patanisha application
leverages on the readily available and accessible mobile
devices and internet technology as opposed to dependency on
desktop computers.
Additionally, unlike some of the legacy systems, the new
application gives the public and in some cases even the lost
person a chance to report themselves in case they are in the
position to. Some legacy systems only allow the administrator

www.ijcat.com 511
to enter the records of missing persons and this level of
bureaucracy and limit of accessibility rights becomes a
hindrance to the reporting of some missing cases.
The new system is also cost effective to develop and maintain
as it does not involve much resources. It does not require the
rather expensive hardware installation. Since its also based on
an emphasis of good will from the public in the reporting and
updates of missing cases, it does not require much
administrative resources e.g human resources.
5. CONCLUSION
These findings are consistent with other studies. A significant
number of missing people has been traced in developed
countries like the United States in the last 3 years by using
various technology based solutions like social media and
personal phone location applications. Kenya is slowly
adopting this strategy but there is need to do better. These
results should be a wakeup call for us to embrace the readily
available technology resources in solving our own problems.
The project was indeed a good opportunity to unveil what an
innovation using the readily available and widely accepted
mobile technology and the internet can achieve. Literature
cited alludes that there exists a gap in prompt reporting,
location and identification of missing people in this country.
Indeed this research comes in handy as a technology that will
allow for timely reporting and identification of missing
people.
Results from the evaluations carried out verify that once a
case has been reported, there is 72 % chance that the victim
will be found. This is a significant improvement from the 30
% probability experienced by use of the old manual system.
The ultimate objective is to reunite friends and family who
have been separated by natural disasters or other reasons.
Evaluations carried out to measure the success level of users
accessing the database returned positive results. Users could
access the mobile application, register missing persons, report
found persons and search for their loved ones.
It is recommended that the mobile phone based application
that has been developed as a prototype should be optimized
more and be adopted to locate actual missing persons. The
application will then give an opportunity to friends and family
members to report missing cases, where the public can view
and revert with leading information regarding the missing
people. In this manner, more families will be reunited with
their missing loved ones and never have to worry again.
The state humanitarian agencies should embrace and promote
this system. This will increase publication of information
about unidentified people and remains, enlisting the public to
help maximise the chances of identification. This way,
families and friends of the missing people will be empowered
to play an active part in searching for their loved ones and
bring vital closure if they are identified. The application will
also feature general information around missing person
investigations and is intended to be a valuable resource if well
tapped on.
6. ACKNOWLEDGMENTS
I express my sincere gratitude to everyone who supported me
throughout the course of this MSC project. I am thankful to
God for using them to grant me inspiring guidance during the
project work. I am sincerely grateful to them for sharing their
truthful and illuminating views on a number of issues related
to the research project.
7. REFERENCES
[1] Paulides, D. (2014). The missing cases:411 Series. 1st
ed. New York: International Publishers.
[2] Smith, W. (2000). Review of national missing persons
agencies.Compass Partnership.
[3] Nina, A. and Fiona, D. (2011). Handbook to Practical
Disaster Preparedness for the Family. 2nd ed. London:
CreateSpace Independent Publishing Platform
[4] Skinner, R. (2010). The missing link to missing people.
1st ed. New York: HarperCollins Publishers.
[5] Lundin, C. (2007). When All Hell Breaks Loose: Stuff
You Need To Survive When Disaster Strikes. 1st ed.
London: Gibbs Smith.
[6] Damon, P. (2006). Introduction to International Disaster
Management. 1st ed. London: Butterworth-Heinemann.
[7] Andy, C. (2010). Using Google's Haiti Missing Persons
Widget, National Public Radio.
[8] Samarajiva, R. (2005). National Early Warning
System.LIRNEasia, [Online]. 2, 2. Available
at:http://lirneasia.net/2005/03/national-early-warning-
system/[Accessed 03 July 2014].
[9] Acharya, M. (2005). Amateur Radio, A potential tool in
emergency operations. 1st ed. New Delhi: A.P.H.
Publishing Corporation.
[10] Levinson, J. and Domb, A. (2013). Disaster Victim
Identification & Privacy. 1st ed. Jerusalem: -The Hebrew
University of Jerusalem.
[11] Creswell, J.W. (2003). Research Design: Qualitative,
Quantitative and Mixed Methods Approaches, 2nd
Ed,
London: Sage Publications.
[12] David M, (2011). Evaluation: From Precision, Recall
and F-Measure to ROC, Informedness, Markedness &
Correlation. Journal of Machine Learning Technologies.
2 (1), 37–63.
[13] Bryman, A. (2008). Social Research Methods, 3rd
Ed,
Oxford: Oxford University Press.

Using a Mobile Based Web Service to Search for Missing People – A Case Study Of Kenya

More Related Content

What's hot (19)

Similar to Using a Mobile Based Web Service to Search for Missing People – A Case Study Of Kenya (20)

More from Editor IJCATR (20)

Recently uploaded (20)

Using a Mobile Based Web Service to Search for Missing People – A Case Study Of Kenya