Telecommunications and electricity consumption in the Province of Trentino, Italy
1. Introduction
Communication technology systems play an important role for nowadays’ technological era. Inventions and advancements focusing in this context have been developed rapidly over the period. Additionally, the universal operation of ICT is also increasing steadfastly during the past few decades (Fettweis & Zimmermann, 2008). One of the most fundamental examples is the telecommunication activities. Nowadays, traditional communication activities have shifted into conventional techniques of communication such as using personal mobile phone and internet. In today’s digital period, majority of the population communicates trough calls, SMSs or emails instead of exchanging written letters. Another evolution is in how people interact with each other through social media that enables them to constantly update and post about their daily lives.
Telecommunications have become one of the most vital aspects for individual living in the present days. The reason is because it helps people to communicate easily with others whether it is for personal or business related. As a result, the development of telecommunication infrastructures occurred in almost every nation. Conversely, the operation for both telecommunication activities and infrastructures require a large amount of energy. Therefore, the rapid development of telecommunications requires high volume of energy and electricity.
The objective of this study is to investigate empirically the relationship between electricity consumption and telecommunication activities in the Province of Trentino. Such knowledge can be beneficial to give insights about the growth and development that occurs within a specific region. In addition, it is also advantageous for government or private institutions to identify the location that needs a construction of power plant. As most of the regions in Trentino consist of mountainous area and small cities, the development of new power plant is estimated to promote the economic growth and expansion of the Province of Trento. (number of electricity power plant in Trentino, the current condition, is it enough for the current demand in the city, does it needed to build a new power plant, how many power plant have they build in the last years)
This paper is structured in the following manner: a literature review and hypothesis development is provided in Section 2. Subsequently, the methodology of this study is elaborated in Section 3. The description of both electricity and telecommunications dataset are produced in Section 4 followed by the analysis and visualisations of the result are examined in Section 5. Finally, conclusion of the study is discussed in Section 6.
2. Literature review
The population of earth is constantly rising over the last century; the approximate amount has exceeded seven billion people at the end of 2017. In result, the increasing population requires more supply of electricity to carry out their daily activities. Electricity is one of the staple needs of human being for power generation in household, office, factory and many others. Moreover, electricity is also very crucial for the economic development of a city or country. A large body of literature has been formed about electricity consumption which includes electricity consumption and economic growth (Ghosh, 2002; Shiu & Lam, 2004; Altinay & Karagol, 2005) electricity consumption in office buildings (Kawamoto, Koomey, & Meier, 2001), energy consumption in information and communication technology (Fettweis & Zimmermann, 2008).
According to Fettweis & Zimmermann (2008), the massive increase in communications technology caused an increasing power requirements to meet the energy demand. This paper will try to forecast the electricity consumption based on telecommunication activities in the Province of Trentino. For the last several years, technology has transformed the way people communicate to one and another. People can easily talk to their friends and families in other parts of the world via calls, SMS or emails. Correspondingly, the rising trend for using electronic equipment to communicate has also caused the increasing electricity consumption on earth (GSMA, 2017).
Figure 1: Total mobile subscribers by region
Figure 2 below shows the total mobile subscribers and mobile internet users worldwide for 2017 and the forecasts for 2025. It can be seen, the total mobile subscribers and mobile internet users have reached 5 billion and 3 billion respectively in 2017. It is predicted in 2025 that the mobile subscribers will be 5.9 billion at an increasing rate of 2.1% and internet users will reach 5 billion with a growing rate of 5.3% (GSMA, 2018).
Figure 2: Total mobile subscribers and internet users worldwide
As a result, the continuous growth of mobile subscribers and internet users initiated the necessities to transfer massive volume of data to user within an applicable duration. The number of data transmission for internet and mobile networks are rising immensely by approximately to a factor of ten each five years (Fettweis & Zimmermann, 2008). Accordingly, the electricity consumption is increasing up to 20% annually and the figure is doubling in nearly 5 years. Furthermore, server farms also require around 108 billion kWh of electricity every year which sums up to more than 1% of the total electricity consumption globally.
Figure 3 below illustrates the increase of electricity consumption in data centres during 2000 and 2005. According to Koomey (2007), the tremendous increase in electricity consumption is seen to be as much as almost 20% every year, equivalent to doubling in 5 years. Generally, server farms require around one 108 billion kWh every year which is equivalent to 1% of power consumption in the world. In addition, the mobile network provider requires roughly 60 billion kWh each year, similar to about 0.33% of electricity consumption worldwide. In aggregate, the total consumption of telecommunications infrastructure is more than 3% of the electricity consumption on earth.
Figure 3: Electricity consumption in data centres
Therefore, this paper will examine the relationship between telecommunication activities and electricity consumption. The study is exploring dataset accessible from a paper titled “A multi-source dataset of urban life in the city of Milan and the Province of Trentino.” Several datasets are released including precipitation, media, telecommunications, administrative region and electrical energy consumption. However, this study will focus on the electrical energy consumption and telecommunications data for the Province of Trentino.
As stated in Fettweis & Zimmermann (2008), the trend of electricity consumption is rising in accordance to the growing requirement of telecommunications. The study will analyse how telecommunication activities influence electricity consumption focusing in the Province of Trentino. The telecommunications data that will be analysed consist of calls out, SMS out, and internet traffic. Subsequently, the hypothesis of this study is stated as there is a relationship between telecommunication activities and electricity consumption. The hypothesis is presented as follows:
Ha: There is a relationship between electricity consumption and telecommunication activities.
3. Methodology
The process and methods in examining the data will be explained in the following section. There are several steps required in conducting the study, the process of the study is as follows:
1. Data acquisition
First of all, the data for this study is retrieved from an open multisource dataset in a study titled “A multi-source dataset of urban life in the city of Milan and the Province of Trentino.” This study is using the electricity data in the Province of Trentino published by SET Distribuzione SPA. Moreover, the study selected 5 random days for telecommunications data in the Province of Trentino issued by Telecom Italia, one of the biggest telecommunication companies in Italy. The entire dataset used in this study is available online at the Harvard Dataverse website.
2. Data preparation & exploration
Subsequent to downloading all dataset from the website, the data will be prepared and explored using a statistical software called R. Prior to conducting data exploration, the essential packages should be loaded to the R script. There are various packages used in the study, including readr, Hmisc, stargazer, ggplot2 and many others. The package readr is used to read rectangular data such as csv and to write csv file. Furthermore, the study also utilized Hmisc package that consist numerous tools for data analysis, importing and noting data, creating tables, assigning missing values, recording variables, etc. Moreover, stargazer package is a function that generate regression analysis result in a well-presented table. It also enables the presentation of outcome from several models and summary statistics.
The next step is to load both electricity value and telecommunications dataset. Since the study is using 5 days for the telecommunications dataset, one of the telecommunications data should be uploaded first. The title of each column is firstly added because the telecommunications dataset has no header. Additionally, the time straw is subtracted to minimum and additional column is added to show the hour and minute of each observation. Thereafter, load the telecommunications data for remaining days and repeat all necessary steps similar to day one. Lastly, all telecommunications data is combined and saved into a new csv file.
3. Descriptive analysis
The third step is to conduct descriptive analysis for the datasets in this study. Descriptive analysis offers an overview of main figures for each variable in the analysis such as the number of observations, number of variables, mean, median, and standard deviation of the data. The descriptive analysis of this study will be further examined in the next section.
4. Regression and analysis
The study will perform Ordinary Least Square (OLS) regression using statistical software called R. The main reason of analysing the data using OLS regression is because of the parameters of the data are unknown and the technique is mostly used in a linear regression model. There are numerous advantages in using OLS regression, one of the benefits is the outcome of OLS regression is relatively easy to understand and high applicability. On the other hand, there are several drawbacks in using OLS regression such as it is sensitive to outliers. There is also a tendency for OLS regression to overfit the data and the technique may not be reliable if the data is not normally distributed.
5. Data visualisation in QGIS
The last phase of the study is data visualisation in QGIS. QGIS is an open source geographic software to “view, edit, and analyse geospatial data”. First of all, a new layer is added to upload the Trentino geojsn data that contains the geospatial data of the province of Trentino. Afterwards, an open layer plug in for Google Map is inserted by selecting the Google Map street option. The next step is to upload the electricity data by creating a layer from a delaminated text file by selecting the file format csv, first record has field name and no geometry. When the electricity file has been uploaded, join the file with Trentino geospatial data by joining the field based on squareid (sqid). In order to show the intensity of electricity in the province of Trentino, the style is changed into graduated, column based on value, and colour red. Finally, the intensity of electricity will be presented with the highest intensity value as the darkest red and lowest intensity as lightest red. Furthermore, repeat the same steps for the telecommunications dataset for the Province of Trentino to visualise the intensity of telecommunications data.
4. Description of data
In this section, the data for this study will be described and the descriptive statistics of the data will be further examined. The first set of data is the electricity dataset for the Province of Trentino with total observation of 2,575. According to Table 1 below, the data consist of two variables; sqid and value.
Table 1: Descriptive analysis of electricity data
1. Sqid
The square id is the “identification string of a given square of the Trentino Grid”. Subsequently, the sqid will be the key variable that will be use to merge the dataset between electricity and telecommunications dataset.
2. Value
This number indicates the ampere value of the electricity that pass through within an area during a specific time. The figure is positive if the direction of the flow is from the national grid into the local site. On the other hand, the figure is negative if the current is going from the local site to national grid. According to the table above, the minimum value is 1 which indicates that there is no negative value and all of the current is flowing from the national grid to local site. The mean for electricity value is 53.881 and the standard deviation is 89.625. The minimum electricity value is 1 and maximum value is 767.
Furthermore, the second dataset for the study is the telecommunications data that consist of random five days’ data. The dates for the month November 2013 are 3rd, 4th and 9th, meanwhile, the dates for the month December 2013 are 14th and 24th. Table 2 below described the key summary of the telecommunications dataset. In total, there are eleven variables in telecommunications data including sqid, TimeSt, ccode, sms_i, sms_o, call_i, call_o and Int, TimeSt, min and hr.
Table 2: Descriptive analysis of telecommunications data
From all variables in telecommunications dataset, 4 variables will be used for the analysis of this study. These variables are sqid, sms_o, call_o and Int.
1. Sqid
Similar to the sqid in the previous electricity dataset in the Province of Trentino, the square id in this dataset is also used to identify the square grid for the Province of Trentino. The total number of observation for square id is 13,494,400.
2. Sms_o
The sms_o is a variable that shows the SMS out activity proportional to the SMS sent within the square id at a specific time. The country where the SMS is sent will be shown in the ccode (country code). In total, 5,117,013 SMS were sent during the observation from the Province of Trentino.
3. Call_o
The explanatory variable call_o variable demonstrates the amount of calls issued in the square id at a particular time. The destination country where the calls made is given in the ccode (country code). In total, 4,979,776 calls were made from the people in Trentino during the observation.
4. Int
The internet traffic activity is an independent variable that represents the amount of CDRs produced inside the square id during a given time. The total observation of internet activity in the data is 6,912,060 times. with the maximum number of 1,969.007.
5. Analysis
• OLS regression
As mentioned in the previous section, the data is analysed using Ordinary Least Squares (OLS) method. The Ordinary Least Squares method is one of the most frequently used techniques in a linear regression model. This method is chosen because the parameters in this study are unspecified. The dependent variable of the model is the value of electricity distribution in the Province of Trentino. On the other hand, the independent variables for this analysis are the telecommunication activities such as SMS out, calls out and Internet traffic.
The result of regression analysis is presented in the table below:
Table 3: Regression result between electricity consumption and telecommunication activities
According to the table above, it can be seen that there is a significant relationship between electricity value and all independent variables with 99% significance level. In result, a very strong relationship exists between electricity consumption and each telecommunication activities such as SMS out, calls out and internet traffic in the province of Trentino. The relationship between the independent and dependent variable will be further discussed in the following section.
The first independent variable is sms_o or SMS out activity. Based on the regression analysis, SMS out activity has a significant influence to the value of electricity at p-value lower than 0.1. This indicates a very strong relationship between both of the independent and dependent variable such as SMS out activity and electricity consumption. Moreover, the relationship between the variables is a positive relationship with coefficient of 0.083. It indicates that there will be a higher intensity of electricity consumption when more people are sending SMS around the Province of Trentino.
The second independent variable is call_o or call out activity. According to the analysis, call out activity has a significant impact to the value of electricity with significance level of 99% which suggests that the relationship is very strong. Additionally, the regression coefficient is 0.270 demonstrating that if more people are making calls from Trentino, the intensity of the electricity value will also be higher. On the other hand, when the intensity of calls out activity is low, the electricity value will be lower.
Finally, the last explanatory variable used in the OLS regression is Int or internet traffic. As seen on the regression analysis, internet traffic also has a significant influence towards electricity value at significance level of 99%. Moreover, the relationship between them is positive with coefficient of 0.091. Nonetheless, the result suggests that when more people are using the internet, the intensity of electricity value will increase too.
Furthermore, the R square of this study is 0.019 and can be interpreted as the independent variables are predicting 1.9% of the dependent variable while the rest remains unknown. This result is relatively low and might be because of lack of other independent variables for predicting electricity consumption. Nonetheless, the result from the OLS regression analysis shows that the independent variables in the model such as SMS out, calls out and internet traffic has significant effect to the electricity usage in the Province of Trentino. Moreover, the independent variables are affecting the value of electricity positively.
• QGIS
Additionally, the data is further analysed using QGIS software to gain an enhance understanding for the visualisation of this dataset. QGIS is an open source topographical information system with various features to analyse spatial data. Please refer to Section 3 to find the detail steps for analysis in QGIS. In this study, QGIS is used to investigate the intensity of several variables in the dataset. This section will show numerous visualisations obtained from QGIS for both electricity and telecommunications data in the Province of Trentino.
o Electricity dataset
The picture below is the visualisation of electricity dataset in the Province of Trentino for the area of Trento (the capital city of Trentino) and other adjacent cities.
Figure 4: QGIS for electricity dataset of Trentino
According to the visualisation, it is seen that the electricity consumption varies over the region. The box with darkest red shows the area with highest electricity consumption and box with lightest colour indicates area with lowest intensity of electricity. Generally, the regions with high intensity of electricity are cities and city centres where a large amount of energy is needed to support the resident’s daily activities. For instance, grid number 5083 and 5084 as shown with arrow in Figure 4 has a very high intensity of electricity consumption. The main reason is because this area is located near Muse, a well-known science museum in Trento. In addition, Figure 5 is a Google Maps screenshot where it exhibits that this region comprises various shops, cafes and restaurants. As a result, this area is consuming a vast amount of electricity to support the businesses.
Figure 5: Google Maps screenshot of area around Muse
o SMS out
Figure X below illustrates the intensity of SMS out for the area around Trento. Similar to the previous visualisation, the darkest colour shows area with the highest intensity for SMS out and the lightest colour shows area with the lowest intensity for SMS out. The area with highest intensity of SMS out will be reveal by circles as shown in Figure 6 and the comparison to Electricity data is presented in Figure 7.
Figure 6: QGIS for SMS out of Trentino
Figure 7: QGIS for electricity & SMS out of Trentino
According to the visualisation in QGIS, three areas are discovered containing the highest intensity of SMS out. The first area with high level of SMS out is around grid 5202 and 5203, it is the location of a city on the eastern side of Trento called Povo. This is the location for scientific and engineering school of the University of Trento. In addition, numerous other research centres are also residing in Povo. The high intensity of SMS is most likely sent by students, staffs, researchers and people residing in Povo. The second area with high intensity of SMS is across grid 5551 and 5668, the location of Gardolo and Melta which is situated on the north of Trento. Originally, Gardolo is known as an agricultural hub for fruits and vegetables. However, the city has been rapidly developed into the manufacturing, trading and housing area of Trento due to its strategic location. In addition, Melta is a city located on the foot of Mt. Calisio and it is exerted as dominion for the nearby towns. Likewise, Melta has also undergone development for residential area and large park of Melta as the venue for many vital sports centre such as football and basketball. Lastly, the third location is around grid 5195 and 5196 where Sopramonte is located. The district is situated approximately 7 kilometres from the capital city of Trento. As seen on Figure 6 and Figure 7, the majority of area that has a high intensity of SMS out activity also has a high concentration of electricity usage which is align with the hypothesis of this study.
o Call out
In Figure 8 below, the intensity of calls out activity around Trento is presented. Figure 9 shows the electricity consumption of Trento for the comparison to calls out activity.
Figure 8: QGIS for calls out of Trentino
Figure 9: QGIS for electricity & calls out of Trentino
In general, there are three areas that has the highest level of intensity for calls out. The first spot is around grid 5551 which is the area around other cities on the north of Trento known as Gardolo and Montevaccino. As mentioned before, Gardolo has been developed into industrial district of Trento. On the one hand, Montevaccino is also transformed to be the residential area of Trento. The two main districts in this region are Montevaccino di Sotto and Montevaccino di Sopra. Additionally, the other site with high concentration of calls out is across grid 5440 until 5441, it is the location of a city on the northeast of Trento called Civezanno. There are various crucial complexes in this area including Telvana Castle that act as the base for local borough. Moreover, the area with high intensity of calls out on the west side of Trento is around the city of Sopramonte. In most cases, the areas that have a high intensity of calls out also have a high intensity of electricity consumption.
o Internet traffic
The image below shows the visualisation of internet traffic for the telecommunications data for the province of Trentino.
Figure 10: QGIS for internet Traffic of Trentino
Figure 11: QGIS for electricity & internet traffic of Trentino
According to Figure 10, there are two main areas with high intensity of internet traffic around Trento. The area with a very strong intensity of internet traffic activity is around grid 5202. This is the location of Povo and the visualisation suggests that a large number of people are using internet in this city as indicated with several dark red boxes surrounding the district. Furthermore, the second area with the highest intensity of internet traffic is around grid 5785 where a city in the north of Trento called Gardolo is located. As shown in Figure 10, dark red boxes are also present suggesting a very strong internet traffic around this region. Figure 11 showed that the majority area with high internet traffic also has a high intensity of electricity consumption which is in accordance with the hypothesis of this study.
In conclusion, the cities with high telecommunications data are mostly similar across the 3 variables consisting of SMS out, calls out and internet traffic. According to the visualisation in QGIS, the majority regions that have a high level of SMS out are Povo, Gardolo, Melta and Sopramonte. Meanwhile, the areas with high intensity of calls out are Gardolo, Montevaccino, Civezanno and Sopramonte. Furthermore, the locations with high concentration of internet traffic activity are Povo and Gardolo.
Moreover, the areas with high intensity of telecommunication activities are the cities located within few kilometres from Trento, the capital city of Province of Trentino. In most cases, these cities have been developed into an industrial, commercial and residential area that integrates directly to Trento. Additionally, some of the regions also exercised as the local municipalities for nearby smaller villages. According to the visualisation, it suggests that there are at least two or more districts with high intensity of telecommunication activities for each category. Generally, most of the areas with high intensity of telecommunications activities such as calls out, SMS out, and internet traffic also has a high intensity of electricity consumption as displayed in above figures. This notion is consistent to the hypothesis of the study; there is a relationship between electricity consumption and telecommunications activity.
6. Conclusion
In today’s modern era, telecommunication activities play a crucial role in the society. The development of various telecommunication methods enables individuals to connect with each other effortlessly without the hindrance of time and distance. People can instantly send text messages and emails to their families and friends in other part of the world. Fascinatingly, technology also allows us to see other people’s face while calling them as if we are talking face to face. Moreover, mobile phones have become a basic necessity for most people around the world. The number of mobile phone user has also grown rapidly over the last decades indicating the shifting of traditional communication method to the conventional communication method.
However, the telecommunication systems require a large amount of energy. This increase of telecommunication trends demands a higher amount of electricity. The purpose of this paper is to investigate the relationship of electricity consumption and telecommunication activities in the Province of Trentino, Italy. In this dataset, the telecommunication activities include SMS out, calls out and internet traffic within this area. The result revealed the existence of statistically significant relationship between telecommunications activities such as SMS out, calls out, and internet traffic with electricity consumption.
To further explore this outcome, the data is illustrated in QGIS to provide visualisation for the intensity of each activities. As a result, the visualisation in QGIS showed that the areas with high intensity of electricity consumption and telecommunication activities are mostly similar across all categories. These regions are predominantly located a few kilometres from the capital city of Province of Trentino commonly known as Trento. In addition, the regions have also been transformed into industrial and residential area to support Trento. According to the analysis, it is discovered that areas with high intensity of telecommunication activities also has a high electricity consumption. This statement is coherent with the hypothesis of this paper as there is a relationship between electricity consumption and telecommunications activity.
References
Altinay, G., & Karagol, E. (2005, November). Electricity consumption and economic growth: Evidence from Turkey. Energy Economics, 27(6), 849-856.
Fettweis, G., & Zimmermann, E. (2008). ICT Energy Consumption – Trends and Challenges. The 11th International Symposium on Wireless Personal Multimedia Communications (WPMC 2008). Dresden: Vodafone Chair Mobile Communications Systems.
Ghosh, S. (2002). Electricity consumption and economic growth in India. Energy Policy, 125-129.
GSMA. (2017). The Mobile Economy 2017. London: GSMA.
GSMA. (2018). The Mobile Economy 2018. London: GSMA.
Huurdeman, A. A. (2003). The Worldwide History of Telecommunications. John Wiley & Sons.
Kawamoto, K., Koomey, J. G., & Meier, A. K. (2001). Electricity Used by Office Equipment and Network Equipment in the U.S.: Detailed Report and Appendices. Ernest Orlando Lawrence Berkeley National Laboratory, Energy Analysis Department. Berkeley: University of California.
Koomey, J. G. (2007). Estimating Total Power Consumption by Servers in the U.S. and the World.
Shiu, A., & Lam, L. P. (2004). Electricity consumption and economic growth in China . Energy Policy, 47-54.
Appendices
##### MILAN / TRENTINO
rm(list=ls())
##### packages
library(readr)
library(Hmisc)
library(stargazer)
library(ggplot2)
library(plyr)
library(lattice)
library(dummies)
library(lfe)
library(lmtest)
##### load data
#### load electricity data for trentino
elecT<- read_csv("Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/electricity line new.csv")
elecTd <-data.frame(elecT)
rm(elecT)
View(electd)
stargazer(elecTd,type="text")
summary(elecTd)
describe(elecTd)
#### load calls data for trentino 1
comT<- read.delim("~/Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/sms-call-internet-tn-2013-11-03.txt", header=FALSE)
comTd <-data.frame(comT)
rm(comT)
View(comTd)
comTd<-rename(comTd,c("V1"="sqid","V2"="TimeSt","V3"="ccode","V4"="sms_i","V5"="sms_o","V6"="call_i","V7"="call_o","V8"="Int"))
stargazer(comTd,type="text")
summary(comTd)
describe(comTd)
## subtract the minimum, so the minimum is zero
comTd<-rename(comTd,c("TimeSt"="TimeStraw"))
comTd$TimeSt<-comTd$TimeStraw
comTd$TimeSt<-comTd$TimeSt-1383433200000
comTd$TimeSt<-comTd$TimeSt/60000
comTd$TimeSt<-comTd$TimeSt/60
comTd$min<-round(floor(( comTd$TimeSt %% 1)*100)*(6/10))
comTd$hr<-trunc(comTd$TimeSt)
comTd$TimeSt<-comTd$hr+comTd$min
comTd$TimeSt<-comTd$TimeSt+201311030000
View(comTd)
stargazer(comTd,type="text")
#### load calls data for trentino 2
comT1<- read.delim("~/Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/sms-call-internet-tn-2013-12-24.txt", header=FALSE)
comTd1 <-data.frame(comT1)
rm(comT1)
View(comTd1)
comTd1<-rename(comTd1,c("V1"="sqid","V2"="TimeSt","V3"="ccode","V4"="sms_i","V5"="sms_o","V6"="call_i","V7"="call_o","V8"="Int"))
stargazer(comTd1,type="text")
summary(comTd1)
describe(comTd1)
## subtract the minimum, so the minimum is zero
comTd1<-rename(comTd1,c("TimeSt"="TimeStraw"))
comTd1$TimeSt<-comTd1$TimeStraw
comTd1$TimeSt<-comTd1$TimeSt-1387839600000
comTd1$TimeSt<-comTd1$TimeSt/60000
comTd1$TimeSt<-comTd1$TimeSt/60
comTd1$min<-round(floor(( comTd1$TimeSt %% 1)*100)*(6/10))
comTd1$hr<-trunc(comTd1$TimeSt)
comTd1$TimeSt<-comTd1$hr+comTd1$min
comTd1$TimeSt<-comTd1$TimeSt+201312240000
View(comTd1)
stargazer(comTd1,type="text")
#### load calls data for trentino 3
comT2<- read.delim("~/Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/sms-call-internet-tn-2013-11-09.txt", header=FALSE)
comTd2 <-data.frame(comT2)
rm(comT2)
View(comTd2)
comTd2<-rename(comTd2,c("V1"="sqid","V2"="TimeSt","V3"="ccode","V4"="sms_i","V5"="sms_o","V6"="call_i","V7"="call_o","V8"="Int"))
stargazer(comTd2,type="text")
summary(comTd2)
describe(comTd2)
## subtract the minimum, so the minimum is zero
comTd2<-rename(comTd2,c("TimeSt"="TimeStraw"))
comTd2$TimeSt<-comTd2$TimeStraw
comTd2$TimeSt<-comTd2$TimeSt-1383951600000
comTd2$TimeSt<-comTd2$TimeSt/60000
comTd2$TimeSt<-comTd2$TimeSt/60
comTd2$min<-round(floor(( comTd2$TimeSt %% 1)*100)*(6/10))
comTd2$hr<-trunc(comTd2$TimeSt)
comTd2$TimeSt<-comTd2$hr+comTd2$min
comTd2$TimeSt<-comTd2$TimeSt+201311090000
View(comTd2)
stargazer(comTd2,type="text")
#### load calls data for trentino 4
comT3<- read.delim("~/Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/sms-call-internet-tn-2013-11-04.txt", header=FALSE)
comTd3 <-data.frame(comT3)
rm(comT3)
View(comTd3)
comTd3<-rename(comTd3,c("V1"="sqid","V2"="TimeSt","V3"="ccode","V4"="sms_i","V5"="sms_o","V6"="call_i","V7"="call_o","V8"="Int"))
stargazer(comTd3,type="text")
summary(comTd3)
describe(comTd3)
## subtract the minimum, so the minimum is zero
comTd3<-rename(comTd3,c("TimeSt"="TimeStraw"))
comTd3$TimeSt<-comTd3$TimeStraw
comTd3$TimeSt<-comTd3$TimeSt-1383519600000
comTd3$TimeSt<-comTd3$TimeSt/60000
comTd3$TimeSt<-comTd3$TimeSt/60
comTd3$min<-round(floor(( comTd3$TimeSt %% 1)*100)*(6/10))
comTd3$hr<-trunc(comTd3$TimeSt)
comTd3$TimeSt<-comTd3$hr+comTd3$min
comTd3$TimeSt<-comTd3$TimeSt+201311040000
View(comTd3)
stargazer(comTd3,type="text")
#### load calls data for trentino 5
comT4<- read.delim("~/Desktop/QMUL/MASTER CLASS/FINAL ASSIGNMENT/sms-call-internet-tn-2013-12-14.txt", header=FALSE)
comTd4 <-data.frame(comT4)
rm(comT4)
View(comTd4)
comTd4<-rename(comTd4,c("V1"="sqid","V2"="TimeSt","V3"="ccode","V4"="sms_i","V5"="sms_o","V6"="call_i","V7"="call_o","V8"="Int"))
stargazer(comTd4,type="text")
summary(comTd4)
describe(comTd4)
## subtract the minimum, so the minimum is zero
comTd4<-rename(comTd4,c("TimeSt"="TimeStraw"))
comTd4$TimeSt<-comTd4$TimeStraw
comTd4$TimeSt<-comTd4$TimeSt-1386975600000
comTd4$TimeSt<-comTd4$TimeSt/60000
comTd4$TimeSt<-comTd4$TimeSt/60
comTd4$min<-round(floor(( comTd4$TimeSt %% 1)*100)*(6/10))
comTd4$hr<-trunc(comTd4$TimeSt)
comTd4$TimeSt<-comTd4$hr+comTd4$min
comTd4$TimeSt<-comTd4$TimeSt+201312140000
View(comTd4)
stargazer(comTd4,type="text")
### combine data for trentino telecommunications
totalcomm <- rbind(comTd, comTd1, comTd2, comTd3, comTd4)
stargazer(totalcomm,type="text")
View(totalcomm)
describe(totalcomm)
summary(totalcomm)
### save into a new csv file
write.csv(totalcomm, file = "totalcomm5days.csv",row.names=FALSE)
#### LINK the step
names(elecTd)
names(totalcomm)
T_elecTd_com<-merge(elecTd,totalcomm, by=c("sqid"))
stargazer(T_elecTd_com,type="text")
View(T_elecTd_com)
###### Analyse how telecommunication affects electricity
M0<-lm(value ~ sms_o+call_o+Int, data=T_elecTd_com)
stargazer(M0, type="text", align=TRUE)
summary(M0)