SAN D - A Dashboard for Analyzing

Employees’ Social Actions over Google Hangouts

Edvan Soares

, Marcos Eduardo

, Vanilson Bur´egio

, Emir Ugljanin

and Zakaria Maamar

Federal Rural University of Pernambuco, Recife-PE, Brazil

State University of Novi Pazar, Novi Pazar, Serbia

Zayed University, Dubai, U.A.E

Keywords:

Dashboard, Google Hangouts, Social Action and Violation.

Abstract:

This paper presents SAN D standing for S ocial ActioN Dashboard. It reports from different perspectives the

social actions that employees in an enterprise execute over social media with focus on Google Hangouts. This

execution might violate the enterprise’s use policies of social actions forcing decision makers take corrective

measures. SAN D is implemented using different technologies like Spring Boot and AngularJS.

1 INTRODUCTION

In previous works (A, 2015a; A, 2016), we designed

and implemented an approach for deﬁning restrictions

over social actions (e.g., chat, post, and comment)

that enterprises’ employees perform over social me-

dia with emphasis on

Google Hangouts

. The ob-

jective of these restrictions is to ensure proper and

efﬁcient use of social media, since many enterprises

are still unsure about their beneﬁts and return-on-

investment (El-Sayed and Westrup, 2011; Kanaracus,

2015). Tracking the execution of social actions leads

to issuing warnings, when deemed necessary, to those

who violate restrictions like when to engage in a chat,

with whom to chat, and how a long a chat remains

active. In this paper, we pursue the same work by

providing a complete analysis of the social actions

that each employee executes along with the restric-

tions that she could have violated. In fact, we develop

a dashboard for decision makers who can drill into

the details collected during the execution of social ac-

tions. This drilling yields into insights such as the

most active employees on social media and the social

actions that each employee performs along with the

restrictions she has violated.

There is an ongoing debate about the role of social

media in the workplace. On the one hand, pros in-

clude reaching out to more customers and tapping into

social data like feedback on services. On the other

hand, cons include distracting employees and facili-

tating security breaches (Sherry, 2015). Approaches

that educate users on how to use social media efﬁ-

ciently and/or suggest preventiveinstead of corrective

actions, are still a few. Circulating photos on the Web

and posting personal details show the severe damages

that these actions have on people’s lives (Roth, 2010).

Wouldn’t it be better to “control” the actions of cir-

culating and posting before performing them? To the

best of our knowledge, our work on restrictions is one

step towards making users aware of the pitfalls of so-

cial media so they can be held liable if recurrent vio-

lations occur, for example (A, 2015a).

In this paper, we present S AN D, standing for

Social ActioN Dashboard, that provides a summary

of employees’ historical data on chat sessions and vi-

olations of restrictions. The summary yields into use-

ful information for decision makers so they are aware

of who is doing what, when a violation took place,

and what response to a violation was taken. The rest

of this paper is organized as follows. Section 2 is an

overviewof our previous work on restrictions over so-

cial actions. Section 3 details S AN D in terms of ar-

chitecture, implementation technologies, and usage.

Section 4 concludes the paper along with some fu-

ture work.

2 BACKGROUND

This section discusses the context of our research,

which is social enterprise (aka enterprise 2.0). Then,

it presents an overview of our approach for deﬁn-

ing and monitoring restrictions over social actions.

Extensive details on this approach are given in (A,

2015a). According to Global Industry Analysts,

Inc. “The global expenditure on Enterprise Web 2.0

214

Soares, E., Eduardo, M., Burégio, V., Ugljanin, E. and Maamar, Z.

SAN D - A Dashboard for Analyzing Employees’ Social Actions over Google Hangouts.

DOI: 10.5220/0006399302140219

In Proceedings of the 12th International Conference on Software Technologies (ICSOFT 2017), pages 214-219

ISBN: 978-989-758-262-2

is forecast to reach

5.7 billion by 2015, driven by

expanding broadband capabilities, decreasing prices,

improving performance of networks, and the devel-

opment of advanced, highly interactive Web 2.0 ap-

plications” (www.strategyr.com, 2016) and “... the

top 15 Web 2.0 vendors will spend

50 billion in

2015 on servers, networks, and other infrastruc-

ture, up from

38 billion in 2014 and

30 billion in

2013” (www.lightwaveonline.com, 2016).

2.1 Context of Research

The architecture upon which a social enterprise op-

erates is depicted in Fig. 1 showing this enter-

prise’s business and social worlds connecting together

through a meet-in-the-middle platform. This platform

hosts Social Machines (SMs) that act as proxies over

Web 2.0 applications (A, 2015c; Bur´egio et al., 2013).

Fig. 1, also, shows interactions between stakeholders

and the business world and between stakeholders and

the social world.

The business world hosts the enterprise’s busi-

ness processes that consist of tasks connected to each

other through dependencies (e.g., prerequisite and co-

prerequisite). The execution of processes might lead

to interacting with the social world so that for in-

stance, some social actions are triggered while oth-

ers are re-executed or canceled. SMs allow these

interactions to take place. The social world hosts

Web 2.0 applications that the enterprise subscribes to.

Some Web 2.0 applications are internal to the enter-

prise (i.e., locally managed) while others are external

(e.g., Facebook) calling for speciﬁc agreements (kind

of service level agreement) between the enterprise and

Web 2.0 applications’ providers. Last but not least,

the meet-in-the-middle platform supports interactions

between the business and social worlds. In this plat-

form, SMs act as proxies over Web 2.0 applications

so that tasks in the business processes trigger social

actions and track their progress at run-time.

2.2 Google Hangouts in Brief

Google Hangouts (hangouts.google.com/), or

commonly called Hangouts

, is a popular com-

munication platform that allows both individuals

and groups to chat as well as make video calls

with up to 25 users at a time. It combines several

Google’s previous services such as Google Talk,

Google+ Messenger, and Google+ Hangouts. Google

promotes Hangouts as the “future” telephony prod-

uct so it has been enhanced with several voice

capabilities of Google Voice. Hangouts comes

along with Google+ Hangouts API (develop-

ers.google.com/+/hangouts) that allows enterprises

and regular users to create collaborative applica-

tions on top of Google Hangouts. This provides a

means for building new features and customizing

Google Hangouts’s behavior. Applications built

for Hangouts are just like regular Web applica-

tions, developed using HTML and JavaScript, but

enriched with Hangouts real-time functionality of

Google+ Hangouts API.

2.3 Categories of Social Actions

Social actions aim at supporting users reach out to

(unknown) peers (e.g., request friendship) and/or en-

gage (unknown)peers in a collaborativeproduction of

content (e.g., co-author a technical report) (A, 2015a).

In Table 1, social actions fall into one of the following

categories: communication, sharing, and enrichment.

As per our previous work (A, 2015a), we use three

properties to deﬁne a social action:

•

Stakeholders

property refers to those who par-

ticipate in a social action in terms of who initiates

it and who reacts to it. This property is manda-

tory for the social actions (e.g., chat) that require

a ”continuous” presence of all stakeholders during

the execution of these actions.

•

Content

property refers to the object that is made

available for and/or by a social action’s stakehold-

ers. This could be text, image, audio, etc.

•

Tool

property refers to a Web 2.0 application

(e.g., Facebook and Google Talk) that makes a so-

cial action available to the stakeholder(s) for exe-

cution.

2.4 Restrictions Over Social Actions

To deﬁne restrictions, we used Object-Constraint

Language (OCL) (Object Management Group, 2012)

as per the following examples:

1. A participant’s identity should be known

(anonymity not allowed) to the employee at

start-up time.

context: Chat

inv R_1: self.stakeholders -->

forAll (s:Stakeholder | s.Name <> null)

where

self

.stakeholders contains the set of

Stakeholder instances representing the partici-

pants who take part in a chat session. | separates

the object from the condition.

2. No more than n characters per chat message

SAN D - A Dashboard for Analyzing Employees’ Social Actions over Google Hangouts

215

Google+ Hangouts, Facebook,

Twitter, etc.

Social

Machines

SM 

i

Google+

Hangouts

SM 

k

Facebook

BPM platforms

Business process

Business world

t

1

t

2

t

i

Interactions

Social world

Unknown stakeholderKnown stakeholder

BPM: business process management

t: task; SM: social machine

Legend

Meet-in-the-middle

platform

Figure 1: Architecture of the social enterprise (adapted from (A, 2015b)).

Table 1: Representative categories of social actions ((A, 2015a)).

Category Description Examples of social actions

Communication

Includes actions that establish back-and-forth

interactions between users, which should en-

gage them in joint operations

Chat with a user or group of users

Request/Accept a relationship connection with

others (e.g., friend, family, and co-worker)

Poke someone

Send direct messages to a user’s inbox

Sharing

Includes actions that establish one-way inter-

actions and allow to create and edit shared

content and to facilitate this content’s con-

sumption

Co-author a text on a Wiki system

Publish a post on a Blog Web site

Upload a video on a public repository

Share schedules, photos, music ﬁles, or any

other content with friends and other users

Subscribe to an RSS Feed

Enrichment

Includes actions that provide additional

[meta] data on shared content by providing

opinions and/or ranking

Comment a post

Rank/Rate a post, page, video, news, etc.

Tag users’ photos, videos, activities, etc.

Recommend a Web site, book, or other prod-

ucts/services to friends and other users

context: Chat

inv R_3: self.messages() -->

forAll(m:Message | m.Size < n)

where

messages()

is deﬁned as follows:

context: Chat

def: messages(): Set(Message) =

self.content --> select(m:Content

| m --> oclIsKindOf(Message))

where

self.content

contains the set of

Content

instances representing messages exchanged dur-

ing a chat session.

3. The enterprise approved chat-tool should be used

context: Chat

inv R_5: self}.tool -->

oclIsKindOf(Internal)

3 SOCIAL ACTION DASHBOARD

SAN D is a descriptive-analytics system for examin-

ing historical data related to chat sessions, only. The

objectives are to understand how an enterprise’s em-

ployees violate restrictions and to support enterprises

in their decision-making processes. SAN D pro-

cesses raw data and presents them in a user-friendly

easily-interpretable format.

3.1 Architecture

Fig. 2 is S AN D’s architecture that extends the

restriction system’s architecture we developed ini-

tially (A, 2016). The additional components are vio-

lation detector, violation repository, analytics service,

and dashboard interface.

• The violation detector is deployed on the client

side and is in charge of identifying and report-

ICSOFT 2017 - 12th International Conference on Software Technologies

216

Send data

Chat app interface

Google+ Hangouts API (Web 2.0 application)

Customized chat app

Activity log

repository

Update

Web portal

Manager

Employee

Action monitor

Set

Synchronize

Restriction

repository

Parse

Restriction manager

Control social machine

Conﬁguration

interface

Business

world

Meet-in-the-middle platform

Social

world

Dashboard

interface

Violation

repository

Analytics Service

Render

Update

Read

Violation

manager

Violation

detector

Figure 2: SAN D’s architecture.

ing violations of restrictions that take place during

chat sessions. The violation detector blocks chat

sessions and sends data to the meet-in-the-middle

platform when violations occur.

• The violation manager is deployed on the meet-in-

the-middle platform and is responsible for receiv-

ing data on violations and storing these data in the

violations repository.

• The violation repository stores all violations re-

ceived by the violation manager. Examples of vi-

olations include employee using chat application

outside the allowed time-periods or employee en-

gaged in more active chat sessions than allowed.

• The analytics services provides APIs that give de-

cision makers the ability to obtain details on the

violations that took place and run summary re-

ports. These services permit to measure for in-

stance, time spent in chat sessions by employees

and total of violations per type and per employee

as well, and to track violationsunder different per-

spectives/views over time.

• The dashboard interface uses the APIs of the ana-

lytics services to provide a graphical user interface

that renders different charts and statistics on em-

ployees’ historical data on chat sessions and vio-

lations of restrictions.

3.2 Implementation and Usage

We discuss some implementation details of S AN D

along with the technologies used in this implementa-

tion. S AN D is divided into two separate parts: front-

end that renders data in a visualization layer, and

back-end that processes data and makes them avail-

able to the front-end. This strategy of implementa-

tion breaks-up the complexity of the system by uncou-

pling the dependencies between the presentation and

business layers and minimizing the impacts of further

changes on the system as a whole.

In the back-end, we use Spring

Boot (projects.spring.io/spring-boot) to implement

the Analytics Services, which provide REST ser-

vices in charge of processing and returning data in

JSON format. Spring Boot is a Java-based framework

that uses Conﬁguration over Convention (CoC) as

a design paradigm to increase development produc-

tivity by discarding for instance, the use of different

XML-based conﬁgurations. SAN D’s Analytics

Services are connected to the violation and activity log

repositories, where data sent by the client-side chat

tool (A, 2016) is stored.

In the front-end part, two views can be high-

lighted: one to visualize the time spent by users us-

ing the chat tool and the other to visualize the amount

of users’ violations according to the previously estab-

lished restrictions. To visualize some metrics, such as

time spent, S AN D’s uses Keshif (www.keshif.me),

a JavaScript-based tool that allows to obtain insights

from data by providing a dynamic interface for data

exploration. The top chart in Fig. 3 sorts in a de-

scending way users according to the number of uses

of Google Hangouts. The middle chart in Fig. 3 shows

the number of conversations separated by intervals of

the time spent in minutes per conversation. It is possi-

ble to realize, for instance, that conversations with du-

SAN D - A Dashboard for Analyzing Employees’ Social Actions over Google Hangouts

217

Figure 3: SAN D’s time-spent view.

ration between zero and twenty minutes have a larger

number of occurrences. Finally, the bottom chart in

Fig. 3 maps the number of uses of Google Hangouts

by date, where it is possible to see the periods that

have a higher and/or lower number of uses. The in-

formation in the three graphs is interconnected, so

it is possible to ﬁlter or make comparisons among

them. For example, to know the information of a

speciﬁc employee, the user simply selects the desired

employee on the chart that lists all employees and the

values of all other charts will be updated, according to

the selected employee. Comparisons are also a feature

that can be highlighted. For instance, to compare the

data of two employees, the user simply selects them

in the chart, and bars with speciﬁc colors for each se-

lected user will be automatically shown in the other

charts, making it possible to compare data (e.g., time

spent in chat conversations) between the selected em-

ployees.

In addition to Keshif, the front-end part

uses AngularJS (www.angularjs.org) and mor-

ris.js (www.morrisjs.github.io/morris.js/) to create

other charts like the one illustrated in Fig. 4. These

charts permit to visualize employees’ violations in

terms of total amount of violations committed by

all employees and amount of violations of a speciﬁc

user. Violations types include the following:

1. Date range during which users are allowed to use

Google Hangouts.

2. Inactive users to detect those who are not active

during a certain time period which is longer then

deﬁned. This requires stopping the inactive ses-

sions and opening new ones.

3. Maximum number of characters to limit the num-

ber of characters per message.

4. Maximum active session to avoid many concur-

rent sessions.

5. Daily session limit to restrict the maximum num-

ber of sessions per day.

4 CONCLUSION

In this paper, we present the architecture and im-

plementation details of S AN D standing for S ocial

ActioN Dashboard. SAN D permits to analyze and

control employees’ actions over Google Hangout so

that it feeds decision makers with various details like

main types of violations that occurred in the work-

place, time of violation execution and violation ex-

ecutors. These details permit to develop best prac-

tices of using social media taking into account an en-

terprise’s concerns like privacy, conﬁdentiality, and

competitiveness. These details also permit to improve

relationships with customers since social actions will

be “controlled” from a productivity performance. In

term of future work, we would like to examine other

forms of social media like Facebook Messenger and

generate patterns of violations so that preventivemea-

sures are put in place.

ICSOFT 2017 - 12th International Conference on Software Technologies

218

Figure 4: SAN D’s violations view.

REFERENCES

A, B. (2015a). No Tile - Blind Review. In Proceedings

of the 19th International IEEE Enterprise Distributed

Object Computing Conference (EDOC’2015), Ade-

laide, Australia.

A, B. (2015b). No Tile - Blind Review. In Proceedings of

the 24th IEEE International Conference on Enabling

Technologies: Infrastructure for Collaborative Enter-

prises (WETICE’2015), Larnaca, Cyprus.

A, B. (2016). No Tile - Blind Review. In Proceedings

of EDOC Workshops 2016 held in conjunction with

the 20th IEEE International Enterprise Distributed

Object Computing Conference (EDOC’2016), Vi-

enna, Austria.

A, B. (January-February 2015c). No Title - Blind Review.

IEEE Internet Computing, 19(1).

Bur´egio, V. A., Meira, S. L., Rosa, N. S., and Garcia, V. C.

(2013). Moving Towards “Relationship-aware” Ap-

plications and Services: A Social Machine-oriented

Approach. In Proceedings of the EDOC’2013 Work-

shops held in conjunction with the 17th IEEE Interna-

tional Enterprise Distributed Object Computing Con-

ference (EDOCW2013), Vancouver, BC, Canada.

El-Sayed, H. and Westrup, C. (2011). Adopting Enter-

prise Web 2.0 Collaborative Technologies in Business:

the Implications for Management Accountants. Re-

search Executive Summary Series, Manchester Busi-

ness School, 7(4).

Kanaracus, C. (2013 (visited February 2015)). Gartner:

Social Business Software Efforts Largely Unsuc-

cessful, For Now. http://www.infoworld.com/d/

applications/gartner-social-business-software-efforts-

largely-unsuccessful-now-211813.

Object Management Group, . (2012). Object Constraint

Language, Version 2.3.1.

Roth, P. (2010). Data Protection Meets Web 2.0: Two Ships

Passing in the Night. UNSW Law Journal, 33(2).

Sherry, D. (2010 (visited September 2015)). Web 2.0

Security Threats and How to Defend Against Them.

http://searchsecurity.techtarget.com/magazineContent

/Web-20-security-threats-and-how-to-defend-against-

them.

www.lightwaveonline.com (March 2015 (checked

out in May 2016)). Web 2.0 Compa-

nies Likely Shape Optical Network Market.

www.lightwaveonline.com/articles/2015/03/web-20-

companies-likely-shape-optical-network-market-

says-lightcounting.html.

www.strategyr.com (June 2010 (checked out in May 2016)).

Global Enterprise Web 2.0 to Reach US$5.7 Billion

by 2015. www.strategyr.com/PressMCP-6116.asp.

SAN D - A Dashboard for Analyzing Employees’ Social Actions over Google Hangouts

219