SHARING EXPERIENCES WITH
MOBILE PHONE DATA COLLECTION
IN UGANDA
FLAVIA KYEYAGO OUMA
UGANDA BUREAU OF STATISTICS
14th October 2015
REGIONAL WORKSHOP & CONFERENCE ON THE USE OF MOBILE TECHNOLOGIES FOR STATISTICAL PROCESSES; UNITED NATIONS
CONFERENCE CENTER, ADDIS ABABA, ETHOPIA; 13-16 OCTONER 2015
CONTENTS
Introduction
Pre Mobile CIS issues
Design & Methodology
Data collection and Extraction
Lessons
Benefits
Challenges
Conclusions
.
INTRODUCTION
Mobile Data Collection (MDC) - use of mobile phones, tablets or PDAs for data collection.
Many platforms that can be used to design surveys to collect specific data i.e statistical
data, photographs, data from a preselection, voice recordings, GPS coordinates, etc.
Platforms vary in ease of use, cost, and features.
Some requirements that must be defined .
sample sizes, budgets, technology services
data quality requirements.
Variances in the interfaces, server side components like databases, data reporting and
management interfaces and available technology services and infrastructure
Mobile Data Collection Application Trends:
Development of Native applications installed on the data collection device
Use of USSD as the messaging framework to send the data to server via SMS
The use of the browser based software to collect and send data to an Application
server
INTRODUCTION
In 2008 , GOU, started a programme called the Community
Information System (CIS)
The main objective was to
collect Administrative data
empower communities to make informed decisions using
readily available up to date information.
The CIS was first implemented in 2009 in about 50 districts
Multisectoral approach and UBOS was in charge of data
processing
used paper based questionnaire and
a system for data entry was developed
However, there were many challenges experienced that
included technical and non technical issues that led to the
exercise stalling
Pre Mobile CIS issues
Infrastructure limitations no - In 2011, the growing use of mobile
electricity and room at Sub- phones pushed the IT team to
innovate and experiment the use pf
counties
mobile phones on the CIS project
Limited HR for entry even at
-The developed a web based solution
both Sub county & district
level
which could be accessed through the
web browsers that are native on the
Entry required long term
employment not sustainable mobile phone
Data delays and data
obsolete yet wanted real
time data for planning at
that level
lack of integration of the
data
-Was
done with the objective of
introducing the alternative of MDC
-Reduce
on
some
infrastructural limitations
of
the
Design & Methodology
The Web application was designed by the IT Team at UBOS using previous
experience
This web interface is accessed through phones with web browsers.
Why Web - web is ubiquitous in nature and can be accessed by any device,
anywhere, anytime
Scope: 5 Modules with about 25 questionnaires, that included administrative
data on health, education, financial institutions, general operations
Technology and Application: mobile device phones with sim cards, Designed
using HTML5, CSS, PHP and Java Script for the front end & Mysql for the back
end.
Server was configured at UBOS § IT team monitored data transmission,
aggregation and extraction
Design & Methodology
The conceptual stages involved
designing the form,
deploying the Form on the server,
deploying the form on the device,
collecting data, sending data to
the server and
downloading the data from the
server and analyzing the data.
the Client module - functionalities of
getting blank forms from the web server
to a mobile phone and also filling the
forms and sending the forms to the server.
allows for setting logical question flow–
thereby making non-applicable questions
hidden from enumerator,
Administration Module : for data
management , data reports, data
exportation, data visualization
Data collection & Extraction
Testing : 3 Districts (Urban/Rural)
Training : Done at the Sub county level
are able to submit the data and get a notification
Staffing
message that the data has been submitted.
Enumerators – CDOs – Parish and Village
Supervisors – District Planners & Population
Supervisors – UBOS
Rolled out to date in about 12 districts
access
to
the
application
is
done
through the browser, with user name &
Password
No data is stored on the phone.
Set validation checks are programmed into the
system for answers entered ( logic skips)
some data cleaning is already completed due to
these features built into the system
system is real time it allows for prompt review of
data quality and makes auditing much easier.
Data can be exported to different formats: CSV,
Ms Excel
Data is captured via the mobile client
and sent via the internet using mobile
data transmission technologies (edge or
GSM) to a central server at UBOS.
Validation is done on the phone before the data
is sent to the server.
Officers
Once a user has filled in the questionnaire, they
MCIS Project planning
Tasks
Project Planning
Duration
6 months
Proof of Concept (3 districts)
3 weeks
Design & Testing by the UBOS IT team
10 weeks
Deployment and Training
5 days
Data Collection
10 days
Generate Draft Data Collection Report
2 days
Lessons
Piloting and iteration are critical
Security
Decide on the course of actions
target data collection efforts to the needs and
Project planning
The team should plan way in advance in
usage the CIS
eliminated the fears of the government
order to loose any time factors
officials
Technology and Team
Composition of the team ( IT & Statisticians) .
System should be fully developed
before the actual data collection exercise
where possible
Training and Support
.
4 days of In-depth training of enumerators
Data integrity and security
Learning curve
and supervisors (questionnaire/System/Trial )
enumerators using the phone for data entry
and continuous support
For the development team
Benefits/ Results
Reduced time
Faster, received in real time
of data collection impacting on presentation of findings
the combination of Data extraction and data entry Processes
Provision of real time data and improved data monitoring process
Reduced cost
More innovation which has lead to more capacity built and Adoption
More support from management, more awareness, training support
Sustainable system that can obtain data on a regular basis
reduced paper use , storage space and paper waste
.
Challenges
Fears to move from PAPI to CAPI – keep adopting and improving
Lack of Policy on Mobile phone use -
Training the CDOs – slow learning curve, emphasize key point & give support
Internet Connectivity
Poor network coverage
- change sim cards to the network that is available/
adding an offline mode .
Battery life
Phone batteries would not last the whole day
– charge with the local area centres and also some have backups and others would use their
phones.
using the in-built touch keypad
size of keypad especially for a very long questionnaire was seen a problem
Errors
small keys -correcting mistakes -decimal point
Data sharing to other MDAs is not yet very feasible
Conclusions
Policy Issues
With
the
increasing
data
demands, NSOs should put in
place policies that support
mobile phones usage
Budgeting
and planning
such projects is important
Capacity
building
benchmarking
encouraged
for
and
should
Infrastructure issues
Expand the use of Mobile phones to
Push for more support and collaboration from
developing partners and TRIs
Do more research on the best platforms
(Cross sectional and long term surveys)
Distinguish factors responsible for error rates
Measure the CBA by carrying out the same
survey with both Paper & Mobile for
comparison purposes
Data Management issues
Network connectivity shortcomings –
consider using off line platforms
Research on mobile GSM Terminals
that can expand network coverage
(PPPs)
.
Management of the full data production cycle
to dissemination and archiving stages should
considered