Mining Citizen Feedback Data for Enhanced Local Government Decision-Making - Project Overview

www.unglobalpulse.org • info@unglobalpulse.org • 2015 1
MINING CITIZEN FEEDBACK DATA FOR ENHANCED
LOCAL GOVERNMENT DECISION-MAKING
PARTNERS: NTB PROVINCIAL GOVERNMENT, LAPOR! UNIT (OFFICE OF THE PRESIDENT
OF INDONESIA)
PROGRAM AREA: ECONOMIC WELL-BEING
BACKGROUND
Following the Reformasi movement, initiated in 1998,
democratisation and decentralisation are central components of
the Government of Indonesia’s policy platform. With greater
administrative responsibilities and increased democratic
accountability, local governments are looking for methods to collect
and understand citizens’ opinions on public services and local
development.
Two citizen feedback systems exist for local governments; one
managed by the central government and the other initiated by sub-
national administrations. LAPOR! (www.lapor.go.id) is the national
complaint system in Indonesia to which a citizen can report any
complaint in Indonesian, either via SMS or the internet. LAPOR! is
designed to improve accountability and, in time, the quality of
public services. Many local governments operate their own SMS-
based citizen feedback mechanisms alongside LAPOR!.
In addition to formal citizen feedback mechanisms, social media
platforms contain vibrant discussions within and between
communities on issues of concern. An opportunity exists to
supplement formal feedback with the passive feedback contained
within public discourse on Twitter.
For this project, Pulse Lab Jakarta combined data on citizens’
opinions from Twitter, LAPOR! and a local SMS-based feedback
mechanism to provide structured insights for local decision-
makers. Nusa Tenggara Barat (NTB) province, one of Indonesia’s
poorest regions, was the focus of this initiative due to the provincial
administration’s interest in new approaches to public governance
and due to the fact that it had been a major beneficiary of the
Australia Indonesia Partnership for Decentralisation.
DATA ANALYSIS OF MULTIPLE SOURCES OF
CITIZEN FEEDBACK
Three data sources were used for this project: two datasets from
two active citizen feedback platforms (national- and provincial-level
systems) and public posts on Twitter, which passively captures
citizens’ opinions on local issues. When multiple types of data are
combined, for example active and passive feedback from citizens,
it is expected that a fuller picture of public opinion is available to
decision-makers.
In order to prepare the datasets relevant to NTB province, the
geospatial information of each message was used to determine if a
tweet was posted in NTB. The Twitter dataset was further refined
by removing spam and other irrelevant tweets. Following this
process, a combined total of over 92,000 messages were
structured by 1600 filtering rules using over 350 keywords, which
were informed by the ten national development priorities.1
1
The ten national priorities include food sufficiency, energy, maritime
development, infrastructure and transportation, education, health, poverty
alleviation, bureaucratic reform, tourism, and industry.
Electric power continuity is a major development issue in NTB
province. The following example presents different perspectives
on the issue and highlights the value of multiple data sources.
TWITTER “The power outage happened at dawn, electricity was
restored at 2PM and yet now the outage happens again, PLN
[the electricity company in Indonesia] does not understand.”
LAPOR! “The power outage happens six times a day in
Mataram, Nusa Tenggara Barat. Although, the duration of the
power outage is less than a minute, it is really disturbing our
activities and might break our house appliances. Need PT PLN
attention to solve this problem.”
SMS GATEWAY “Why do power outages frequently happen after
the change of the Regent?”
SUMMARY
Like many middle-income countries, decentralisation is a national priority in Indonesia. As their administrative responsibilities and fiscal resources
have increased, local governments have sought better evidence to inform policy. In this project, Pulse Lab Jakarta explored the contribution of
advanced data analytics to local government decision-making by generating insights from a combination of existing complaint systems and passive
feedback from citizens on social media. The results demonstrate the potential utility of (a) near real-time information on public policy issues and
their corresponding locations within defined constituencies, (b) enhanced data analysis for prioritisation and rapid response, and (c) deriving
insights on different aspects of citizen feedback. The publication of citizen feedback on public-facing dashboards can enhance transparency and
help constituents understand how their feedback is processed.
HOW TO CITE THIS DOCUMENT:
UN Global Pulse, “Mining Citizen Feedback Data for Enhanced
Local Government Decision-Making”, Global Pulse Project Series
no.16, 2015.

www.unglobalpulse.org • info@unglobalpulse.org • 2015 2
Twitter LAPOR! SMS
Num. of msg. used 91,995 171 115
Timeline June 2013-Dec 2014 Oct 2013
Table 1: Number of messages used for this project and their date of
submission.
The next step involved the application of several analytical methods
including volume analysis, text analysis, locational analysis, topic
detection and anomaly detection, among others.
INSIGHTS & OUTCOMES
Visualising changing trends in the volume of messages provides an
easy way to understand popular priorities and concerns in near
real-time. In June and July 2013, a spike in comments and
messages about poverty alleviation, in particular regarding the
unequal distribution of a social protection programme, was
identified along with locational information on where citizens raised
the issue (see Figures 1 and 2). The locational information in
particular proved useful to the NTB administration in planning its
response.
Figure 1: Dynamics of the volume of LAPOR! by the 10 national priorities
Figure 2: A visualization based on the ten cities/regencies within the
province (the darker the colour, the higher the density of messages
classified as Poverty Alleviation)
Some issues are inherently coupled, requiring multiple authorities
to investigate and respond. The dashboard used word co-
occurrence analysis to identify linked problems. Some examples of
issue coupling included:
A volume analysis gives a good snapshot of current issues raised
by citizens, but when the volume of messages is high across many
issue areas policy-makers cannot easily determine which issues to
prioritise. A simple but automatic anomaly detection method based
on a dynamic threshold value was applied to the database in order
to identify automatically sudden spikes. For instance, a sudden
volume change on Government Aid (“Bantuan Pemerintah”) was
identified, as shown in Figure 4, enabling prioritisation by decision-
makers.
Figure 4: The dashboard automatically highlights priority issues
CONCLUSIONS
This project demonstrated the potential of using multiple sources
of near real-time information for decision-making in local
government, especially the value of combining active citizen
complaints with passive opinions expressed via social media.
The visualisation of complex datasets enables non-technology
literate officials to process citizens’ feedback at low cost and at
scale, as well as to prioritise trending issues based on enhanced
data analysis.
The project highlights the potential of existing datasets, but also the
need to integrate new information management systems into
national and local governance. It represents another step towards
realising the potential of networked governance in Indonesia.
IMPLICATIONS & RECOMMENDATIONS
• Social media captures a richer array of issues than formal
complaint mechanisms and should be considered an
essential source of citizen feedback.
• It is advisable that provincial administrations consider
integrating real-time data from their formal feedback
systems into the dashboard, as opposed to mining
historical data.
• The integration of additional data sources should also be
considered when designing dashboards.
• Whenever possible, provincial administrations should publish
this type of citizen feedback dashboard in order to enhance
transparency and to help constituents understand how their
feedback is processed.
FOOD SUFFICIENCY AND ENERGY SECURITY (TWITTER)
“The brown rice and broccoli prices go up following the
increase in fuel price”
HEALTH AND POVERTY ALLEVIATION (LAPOR!)
“Dear health minister, I have suffered heart disease for 14
years. Is there any way for poor people to receive free surgery?
Is there any possibility that the surgery will be free in future?”

Mining Citizen Feedback Data for Enhanced Local Government Decision-Making - Project Overview

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Mining Citizen Feedback Data for Enhanced Local Government Decision-Making - Project Overview

Similar to Mining Citizen Feedback Data for Enhanced Local Government Decision-Making - Project Overview (20)

More from UN Global Pulse

More from UN Global Pulse (20)

Recently uploaded

Recently uploaded (20)

Mining Citizen Feedback Data for Enhanced Local Government Decision-Making - Project Overview