Instaling Packages

Loading Data and Preliminary

setwd("C:/Users/ahmet/Desktop/MEF_BDA/R_Data")
data <- read_xlsx("ptf-smf.xlsx")

data2= data %>% select(Tarih,PTF,SMF)
data2
## # A tibble: 720 x 3
##    Tarih                 PTF   SMF
##    <dttm>              <dbl> <dbl>
##  1 2020-09-01 00:00:00  302.  332.
##  2 2020-09-01 01:00:00  300.  325.
##  3 2020-09-01 02:00:00  293.  318.
##  4 2020-09-01 03:00:00  290   320 
##  5 2020-09-01 04:00:00  290   330 
##  6 2020-09-01 05:00:00  290   339 
##  7 2020-09-01 06:00:00  292.  342.
##  8 2020-09-01 07:00:00  295.  345 
##  9 2020-09-01 08:00:00  307.  339.
## 10 2020-09-01 09:00:00  315.  346.
## # ... with 710 more rows
data3=data2 %>% mutate(diff=PTF-SMF, smf_direc=ifelse(PTF>SMF,"Energy Surplus",ifelse(PTF<SMF,"Energy Deficit","Balance"))) %>% 
  select(Tarih,PTF,SMF,diff, smf_direc)
data3
## # A tibble: 720 x 5
##    Tarih                 PTF   SMF  diff smf_direc     
##    <dttm>              <dbl> <dbl> <dbl> <chr>         
##  1 2020-09-01 00:00:00  302.  332. -30   Energy Deficit
##  2 2020-09-01 01:00:00  300.  325. -25   Energy Deficit
##  3 2020-09-01 02:00:00  293.  318. -25   Energy Deficit
##  4 2020-09-01 03:00:00  290   320  -30   Energy Deficit
##  5 2020-09-01 04:00:00  290   330  -40   Energy Deficit
##  6 2020-09-01 05:00:00  290   339  -49   Energy Deficit
##  7 2020-09-01 06:00:00  292.  342. -50   Energy Deficit
##  8 2020-09-01 07:00:00  295.  345  -49.6 Energy Deficit
##  9 2020-09-01 08:00:00  307.  339. -31.5 Energy Deficit
## 10 2020-09-01 09:00:00  315.  346. -31   Energy Deficit
## # ... with 710 more rows

Distribution of Results (Energy Surplus, Energy Deficit, Balance)

data_bar=data3 %>% select(smf_direc)

ggplot(data_bar, aes(smf_direc)) +
  geom_bar(fill = "forest green") + 
  labs(x = "", y = "Count")

Percentage of energy deficit is higher than Enerji surplus and balance

Daily Situation

data3_day=data3 %>% group_by(day = lubridate::day(Tarih)) %>% summarise(total_ptf=sum(PTF), total_smf=sum(SMF))
## `summarise()` ungrouping output (override with `.groups` argument)
ggplot()+
    geom_line(data=data3_day,aes(x=day, y=total_ptf,group=1, colour="darkblue"),size=1 )+
    geom_line(data=data3_day,aes(x=day, y=total_smf,group=1, colour="red"),size=1 )+
    scale_color_discrete(name = "Categories", labels = c("Daily PTF", "Daily SMF")) + labs(x = "day", y = "Amount")

Work Hours Effect

data_work_hours=data3 %>% mutate(hour = lubridate::hour(Tarih), day = lubridate::day(Tarih)) %>%
  select(Tarih,hour,day, PTF, SMF, diff, smf_direc) %>% 
  mutate(work_hour=ifelse(hour>=9 & hour<19 & day%%7!=6 & day%%7!=5,"yes","no")) %>%
  select(Tarih,hour,day, PTF, SMF, diff, smf_direc, work_hour)

data_work = data_work_hours %>% select(everything()) %>%
  filter(work_hour=="yes")


data_holiday = data_work_hours %>% select(everything()) %>%
  filter(work_hour=="no")

Out of the Work

ggplot(data_holiday %>% select(smf_direc), aes(smf_direc)) +
  geom_bar(fill="red") + 
  labs(x = "", y = "Count")

Work

ggplot(data_work %>% select(smf_direc), aes(smf_direc)) +
  geom_bar(fill = 'blue') + 
  labs(x = "", y = "Count")

We can say that in work hours percentage of the energy deficit is too high. Another mean that is higher than normal situation. In result, it is obviously that predicting the real demand is very hard in work hours