Nowadays every person owns a smartphone, which is necessary at working,
in being connected with others and to get information immediately. The
enormous amount of data about the activity of users of mobile networks
means a treasure to telecommunications companies. This kind of big data
can be used for a lot of solutions. For instance, one can compare data about
the positions of users with the data of public transportation. This raising
hides a lot of opportunities in itself, like optimising routes of transportation,
ascertain the density of routes, creating new optimal traffic paths and so on.
During the process of position data, noise can occur which causes outliers in
the outcome. To have correct results, it is essential to filter these inaccuracies.
In my work, I examined different kind of smoothing and filtering
algorithms, which properties were taken into account when I chose the most
appropriate one for my problem. There are numerous tools to solve the inacc-
uracy problem, like regression analysis, kernel-based smoother. Kalman-filter
seemed to be the ideal solution that gives us an optimal estimation of the
state of moving systems according to measured values and possible measure
of errors. After evaluating some types of this method, I have chosen the one
which fitted best the ideal trajectory, and interpreted the results on a real
Storing and processing that massive amount of data is a significant technological challenge.
Open source big tools offer solutions to handle huge datasets. With the help of these technologies
one can build robust, scalable and fault tolerant systems.
After examining some of the popular big data tools, I chose the ideal components
that build up a system to process the incoming cellular data.