International Journal of Advanced Research in
Computer Science & Technology (IJARCST 2017)
41
Vol. 5, Issue 3 (July - Sept. 2017)
ISSN : 2347 - 8446 (Online)
ISSN : 2347 - 9817 (Print)
www.ijarcst.com © All Rights Reserved, IJARCST 2013
A Cloud Based Solution For Predicting Strength Level of
Residential Houses Using Data Analytics
I
Kirema D. Mutembei,
II
Andrew M. Kahonge
I
School of Computing and Informatics, University of Nairobi, Kenya
II
School of Computing and Informatics, University of Nairobi, Kenya
I. Introduction
Kenya and particularly Nairobi, has experienced frequent collapse
of buildings. This frequent collapse of residential houses in
Nairobi has lead to injuries and deaths in the past is a matter of
great concern. The recent collapse being in Kwale - Embakasi,
Nairobi City County. Lately, other buildings collapsed in areas
like Huruma, makongeni in Thika, Roysambu, along Thika road
superhighway just behind TRM mall killing several people and
other scores injured [1]. These accidents are avoidable but still
they have claimed lives of innocent Kenyans, robbing them of
their families and loved ones. The tenants have suffered in such
conditions since they are not informed of the condictions of the
house before occupying them.
Kenyans thus, need to be informed of their safety in the residential
homes they live in. An algorithm will be sort that will utilize the
construction metrics such as grade of concrete, grade of steel used
and the location – wetland, soil type i.e. black cotton or red soil.
These metrics will be analysed and used to compute the strength
level of a given building. It also will rank the building by some star
level to deine if its strong enough and safe for human occupation.
The same algorithm will be used to determine the inssuance
of occupancy certiicates among other factors. This paper has
enhanced ICT usage by developing a prototype for predicting
residential building strength level and ensuring residents are well
informed of the residential home they wish to live in. The authority
will also be informed hence easy to manage the buildings in the
County [2]. This paper sought to study the usage of data analytics
and cloud computing with view to developing a Software as a
Service based residential home classiication service. With among
other objectives as to determine dataset analytics techniques used
in data analysis, develop an algorithm that will analyse the dataset
and provide results based on star rating and to test the algorithm
and evaluate techniques of exposing APIs to developers.
II. Related Studies
The construction industry has been identiied as one area where
application of technology in capturing data for use in analysis
for determining a house strength is less used if not at all. The
dificulties with emerging technologies applications in construction
ields where the metrics used, that could function as valuable
input information on analytic and predictable methods to provide
structural strength level of a building, have not been used [3].
Data mining as a process of identifying interesting patterns
from large database [3] [4]. An iterative and exploratory process
achieved via automated or manual methods. The primary roles
of data mining are: Prediction that involves the use of variables
to predict unknown future events or values of a given outcome
and description that involves patterns identiication that portray
the data used in a signiicant manner. Algorithms are categorized
according to various distinctions. These includes classiication
methods - used to discover predictive relationships for categorical
variables, predictive methods - used to discover predictive
relationships for numeric variables and association methods –
used in association rule discovery[4], he proposed a conceptual
framework for graphical classiication for applications of data
mining techniques to inancial fraud detection.
Prediction and sequential pattern analysis determines the
relationship between dependent and independent variables. This
analysis technique was utilized in buildings to predict the strength
level of a residential building; some metrics like soil type, concrete
grade and steel grade are independent variable while strength level
can be dependent variable. Based on such data, a regression curve
can be used for strength prediction [6].
Cloud computing paradigm - Software as a Service (SaaS), which
this papers’ prototype is based has been used [7]. Cloud Computing
can adapt dynamically to different circumstances, like the number
of requests, making it making it a good choice for the analytical
algorithm deployment. [8-10].
The adoption to use of SaaS for web development [11] and
deployment in today’s technologies with [12], saying it is the
backbone to Service Oriented Architecture (SOA) [13,14].
Cloud computing goal is to share data, computational calculations
and services to the users. The prediction analytical algorithm also
needs to share analysed information to its users [15-17].
The basic conceptual framework for the algorithm is as shown
in igure one below.
Abstract
Kenya has experienced frequent collapse of residential houses leading to injuries and deaths. Most tenants are not aware of the status
of such buildings. An algorithm to classify and rate the houses based on either structural strength was used. Cloud computing has
been used to actualize the algorithm. The algorithm implementation involved irst inding the metrics that determine the strength
level of a residential building. Weighted metrics actualized the algorithm and results were displayed in star rating scale where ive
star rated house meant was strong enough while one star rated residential house meant a weak building and prone to collapse.
Node.js and MongoDB were among technologies used. GPS was implemented to locate exact geolocation building. The prototype
results revealed 97 percent accuracy in predicting the strength level of residential buildings. This was so through real time analysis
of metrics used to predict the strength level.
Keywords
Algorithm, Analytics, Metrics, Geographical Information System (GIS), Cloud computing, Software as a service (SaaS).