API tools faq
paste
Guest User

Untitled

a guest
Jun 21st, 2018
106
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 10.16 KB | None | 0 0
raw download clone embed print report
  1. #IOT Energy Analytics - Review of Submetering
  2. install.packages("dplyr")
  3. library(dplyr)
  4. install.packages(c("tidyr", "devtools"))
  5. library(tidyr)
  6. library(devtools)
  7.  
  8.  
  9. #Data Install
  10. setwd("~/Desktop/IOT Analytics")
  11. library(readr)
  12. household_power_consumption <- read_csv("household_power_consumption.txt")
  13. View(household_power_consumption)
  14. attributes(household_power_consumption)
  15. housevs2 <- read.table("household_power_consumption.txt", header = TRUE, sep = ";", fill = TRUE)
  16. energy_iot <- housevs2
  17. View(energy_iot)
  18. summary(energy_iot)
  19. str(energy_iot)
  20.  
  21. #One column for Date and Time Stamp
  22. energy_combined <-cbind(energy_iot,paste(energy_iot$Date,energy_iot$Time), stringsAsFactors=FALSE)
  23. colnames(energy_combined)[10] <-"DateTime"
  24. View(energy_combined)
  25. energy_combined <- energy_combined[,c(ncol(energy_combined), 1:(ncol(energy_combined)-1))]
  26. head(energy_combined)
  27. View(energy_combined)
  28.  
  29. #Convert Date and Time
  30. energy_combined$DateTime <- strptime(energy_combined$DateTime, "%d/%m/%Y %H:%M:%S")
  31. energy_combined$Date <- as.Date(energy_combined$Date, "%d/%m/%Y")
  32. str(energy_combined)
  33. summary(energy_combined)
  34.  
  35. #Attributes redefined
  36.  
  37. energy_combined$Sub_metering_3 <- as.factor(energy_combined$Sub_metering_3)
  38. str(energy_combined)
  39.  
  40. #Group Data and transform Posix _ dec
  41. library(dplyr)
  42.  
  43. energy_combined$Date <- as.POSIXct(energy_combined$Date)
  44. energy_combined$DateTime <- as.POSIXct(energy_combined$DateTime)
  45.  
  46. #Descriptive Stats
  47. summary(energy_combined)
  48.  
  49. #Feature Engineering
  50. energy_combined$Global_active_power<-NULL
  51. energy_combined$Global_reactive_power<-NULL
  52. energy_combined$Voltage<-NULL
  53. energy_combined$Global_intensity<- NULL
  54.  
  55.  
  56. #Df Quarter1 2007
  57. install.packages("lubridate")
  58. library(lubridate)
  59. dim(energy_combined)
  60. str(energy_combined)
  61.  
  62. #Subset Dataframes 2007 - Submeter 1 - March
  63. DF2007_march <- subset(energy_combined, Date == "2007-03-01"| Date =="2007-03-03"| Date=="2007-03-04"| Date=="2007-03-05"|Date=="2007-03-06"|Date=="2007-03-07"|Date=="2007-03-08"|Date=="2007-03-09"|Date=="2007-03-10")
  64. View(DF2007_march)
  65.  
  66. #Subset Dataframes 2007 - Submeter 2 - August
  67. DF2007_august <- subset(energy_combined, Date == "2007-08-01"| Date =="2007-08-03"| Date=="2007-08-04"| Date=="2007-08-05"|Date=="2007-08-06"|Date=="2007-08-07"|Date=="2007-08-08"|Date=="2007-08-09"|Date=="2007-08-10")
  68. View(DF2007_august)
  69.  
  70. #Subset Dataframes 2007 - Submeter 3 - December
  71. DF2007_dec <- subset(energy_combined, Date == "2007-12-01"| Date =="2007-12-03"| Date=="2007-12-04"| Date=="2007-12-05"|Date=="2007-12-06"|Date=="2007-12-07"|Date=="2007-12-08"|Date=="2007-12-09"|Date=="2007-12-10")
  72. View(DF2007_dec)
  73.  
  74. #Subset Dataframes 2008 - Submeter 1 - march
  75. DF2008_march <- subset(energy_combined, Date == "2008-03-01"| Date =="2008-03-03"| Date=="2008-03-04"| Date=="2008-03-05"|Date=="2008-03-06"|Date=="2008-03-07"|Date=="2008-03-08"|Date=="2008-03-09"|Date=="2008-03-10")
  76. View(DF2008_aug)
  77.  
  78. #Subset Dataframes 2008 - Submeter 2 - august
  79. DF2008_august <- subset(energy_combined, Date == "2008-08-01"| Date =="2008-08-03"| Date=="2008-08-04"| Date=="2008-08-05"|Date=="2008-08-06"|Date=="2008-08-07"|Date=="2008-08-08"|Date=="2008-08-09"|Date=="2008-08-10")
  80. View(DF2008_august)
  81.  
  82. #Subset Dataframes 2008 - Submeter 3 - december
  83. DF2008_dec <- subset(energy_combined, Date == "2008-12-01"| Date =="2008-12-03"| Date=="2008-12-04"| Date=="2008-12-05"|Date=="2008-12-06"|Date=="2008-12-07"|Date=="2008-12-08"|Date=="2008-12-09"|Date=="2008-12-10")
  84. View(DF2008_dec)
  85.  
  86. #Subset Dataframes 2009 - Submeter 1 -march
  87. DF2009_march <- subset(energy_combined, Date == "2009-03-01"| Date =="2009-03-03"| Date=="2009-03-04"| Date=="2009-03-05"|Date=="2009-03-06"|Date=="2009-03-07"|Date=="2009-03-08"|Date=="2009-03-09"|Date=="2009-03-10")
  88. View(DF2009_march)
  89.  
  90. #Subset Dataframes 2009 - Submeter 2 - august
  91. DF2009_august <- subset(energy_combined, Date == "2009-08-01"| Date =="2009-08-03"| Date=="2009-08-04"| Date=="2009-08-05"|Date=="2009-08-06"|Date=="2009-08-07"|Date=="2009-08-08"|Date=="2009-08-09"|Date=="2009-08-10")
  92. View(DF2009_august)
  93.  
  94.  
  95. #Subset Dataframes 2009 - Submeter 3 - december
  96. DF2009_dec <- subset(energy_combined, Date == "2009-12-01"| Date =="2009-12-03"| Date=="2009-12-04"| Date=="2009-12-05"|Date=="2009-12-06"|Date=="2009-12-07"|Date=="2009-12-08"|Date=="2009-12-09"|Date=="2009-12-10")
  97. View(DF2009_dec)
  98.  
  99. #Subset Dataframes 2010 - Submeter 1 - march
  100. DF2010_march <- subset(energy_combined, Date == "2010-03-01"| Date =="2010-03-03"| Date=="2010-03-04"| Date=="2010-03-05"|Date=="2010-03-06"|Date=="2010-03-07"|Date=="2010-03-08"|Date=="2010-03-09"|Date=="2010-03-10")
  101. View(DF2010_march)
  102.  
  103. #Subset Dataframes 2010 - Submeter 2 -august
  104. DF2010_august <- subset(energy_combined, Date == "2010-08-01"| Date =="2010-08-03"| Date=="2010-08-04"| Date=="2010-08-05"|Date=="2010-08-06"|Date=="2010-08-07"|Date=="2010-08-08"|Date=="2010-08-09"|Date=="2010-08-10")
  105. View(DF2010_august)
  106.  
  107. #Subset Dataframes 2010 - Submeter 3 -november
  108. DF2010_nov <- subset(energy_combined, Date == "2010-11-01"| Date =="2010-11-03"| Date=="2010-11-04"| Date=="2010-11-05"|Date=="2010-11-06"|Date=="2010-11-07"|Date=="2010-11-08"|Date=="2010-11-09"|Date=="2010-11-10")
  109. View(DF2010_nov)
  110.  
  111. #Remove variable noise for august
  112. DF2007_august$Sub_metering_2 <- NULL
  113. DF2007_august$Sub_metering_3 <- NULL
  114. DF2007_august$DateTime <- NULL
  115. DF2007_august$Global_intensity <- NULL
  116. DF2007_august$Global_active_power <- NULL
  117. DF2007_august$Global_reactive_power <- NULL
  118. DF2007_august$Date <- NULL
  119. DF2007_august$Time<- NULL
  120. View(DF2007_august)
  121.  
  122. str(DF2007_august)
  123. glimpse(DF2007_august)
  124.  
  125.  
  126. library("")
  127. ?tslm
  128. ?forecast
  129.  
  130.  
  131. #forecast for august 2007_ten days in august_submeter1
  132. Fore2007 <- ts(DF2007_august)
  133. plot.ts(Fore2007)
  134. plot(Fore2007)
  135. plot.forecast(Fore2007)
  136. View (Fore2007)
  137. summary(Fore2007)
  138. Augustfore_2007 <- forecast()
  139. Augustfore_2007
  140. augustoutput <- as.data.frame(Augustfore_2007)
  141. write_excel_csv(augustoutput, "August2007.csv")
  142.  
  143. #forecast focusing on Submeter 1 for August 2007
  144.  
  145. DF2007_august$Sub_metering_2 <- NULL
  146. DF2007_august$Sub_metering_3 <- NULL
  147. View(Fore2007)
  148. plot(Fore2007)
  149. Aug2007_sub1 <- forecast(Fore2007, level = c(80,95), lambda = NULL)
  150. Aug2007_sub1
  151. Augsub1 <- as.data.frame(Aug2007_sub1)
  152. write_excel_csv(Augsub1, "augsub1.csv")
  153.  
  154. #Voltage August 2007
  155. plot(Augustfore_2007)
  156.  
  157.  
  158. #Forecast time series #2 (Dec_2008_winter)
  159. DF2008_dec$DateTime<-NULL
  160. DF2008_dec$Date<-NULL
  161. Fore2008 <- ts(DF2008_dec)
  162. is.na(Fore2008)
  163. plot(Fore2008)
  164. View(Fore2008)
  165. Decfore_2008 <- forecast(Fore2008, level = c(80,95), lambda = NULL)
  166. Decfore_2008
  167. plot(Decfore_2008)
  168. decoutput <- as.data.frame(Decfore_2008)
  169. write_excel_csv(decoutput, "Dec2008.csv")
  170.  
  171. #Dec 2008 focus on submeter 2
  172. DF2008_dec$Sub_metering_3<- NULL
  173. DF2008_dec$Sub_metering_1 <- NULL
  174. dec2008_sub2 <- forecast(Fore2008, level = c(80,95), lambda = NULL)
  175. plot(dec2008_sub2)
  176. dec2008_sub2
  177. decsub2<-as.data.frame(dec2008_sub2)
  178. write_excel_csv(decsub2,"dec2008sub2.csv")
  179.  
  180. #Voltage dec 2008 sub 2
  181. DF2008_dec$Global_active_power <-NULL
  182. DF2008_dec$Global_reactive_power <- NULL
  183. DF2008_dec$Global_intensity <- NULL
  184. DF2008_dec$Time<- NULL
  185. Fore2008 <- ts(DF2008_dec)
  186. plot(Fore2008)
  187.  
  188.  
  189. #Forecast time series #3 (March 2009_spring)
  190. DF2009_march$DateTime <- NULL
  191. DF2009_march$Date<-NULL
  192. DF2009_march$Time<-NULL
  193. Fore2009 <- ts(DF2009_march)
  194. View(Fore2009)
  195. plot(Fore2009)
  196.  
  197.  
  198. Marchfore_2009 <- forecast(Fore2009, level = c(80,95), lambda = NULL)
  199. Marchfore_2009
  200. marchustoutput <- as.data.frame(Marchfore_2009)
  201. write_excel_csv(marchustoutput, "March 2009.csv")
  202.  
  203. #March 2009 focus on submeter 3
  204. DF2009_march$Sub_metering_1<-NULL
  205. DF2009_march$Sub_metering_2<-NULL
  206. MAR2009_sub3 <- forecast(Fore2009, level = c(80,95), lambda = NULL)
  207. plot(MAR2009_sub3)
  208. dec2008_sub2
  209. marsub3<-as.data.frame(MAR2009_sub3)
  210. write_excel_csv(marsub3,"mar2009sub3.csv")
  211.  
  212. #Voltage Macrch 2009 sub3
  213.  
  214. DF2009_march$Global_active_power <-NULL
  215. DF2009_march$Global_reactive_power <- NULL
  216. DF2009_march$Global_intensity <- NULL
  217.  
  218. Fore2009 <- ts(DF2009_march)
  219. plot(Fore2009)
  220.  
  221.  
  222.  
  223.  
  224. #Decomposing Seasonal Trends for Energy data _ August 2007
  225. library("dplyr")
  226.  
  227. householdtimeseries_2007_activepower<- ts(energy_combined$Global_active_power, frequency = 12, start = c(2007,1), end = c(2007,12))
  228. householdtimeseries_2007
  229.  
  230.  
  231.  
  232.  
  233.  
  234.  
  235.  
  236.  
  237.  
  238. plot.ts(householdtimeseries_2007)
  239. season2007<-decompose(householdtimeseries_2007_activepower, "multiplicative")
  240. season2007
  241. plot(as.ts(season2007$seasonal))
  242. plot(as.ts(season2007$random))
  243. plot(as.ts(season2007$trend))
  244. str(season2007)
  245. summary(season2007)
  246. season2007_df <- data.frame(season2007 = c(season2007), time = c(time(season2007)))
  247. write_excel_csv(season2007_df,"season2007.csv")
  248.  
  249.  
  250. #Decomposing Seasonal Trends for Energy data _ Dec 2008
  251.  
  252. library("dplyr")
  253. householdtimeseries_2008 <- ts(energy_combined, frequency = 4, start = c(2008,1), end = c(2008,12))
  254. householdtimeseries_2008
  255.  
  256. plot.ts(householdtimeseries_2008)
  257. season2008<-decompose(householdtimeseries_2008, "multiplicative")
  258. season2008
  259. plot(as.ts(season2008$seasonal))
  260. plot(as.ts(season2008$random))
  261. plot(as.ts(season2008$trend))
  262. season2008_df <- data.frame(season2008 = c(season2008), time = c(time(season2008)))
  263. write_excel_csv(season2008_df,"season2008.csv")
  264.  
  265.  
  266. #Decomposing Seasonal Trends for Energy data _ March 2009
  267.  
  268.  
  269. library("dplyr")
  270. householdtimeseries_2009 <- ts(energy_combined, frequency = 4, start = c(2009,1), end = c(2009,12))
  271. householdtimeseries_2009
  272.  
  273. plot.ts(householdtimeseries_2009)
  274. season2009<-decompose(, "multiplicative")
  275. season2009
  276. plot(as.ts(season2009$seasonal))
  277. plot(as.ts(season2009$random))
  278. plot(as.ts(season2009$trend))
  279. season2009_df <- data.frame(season2009 = c(season2009), time = c(time(season2009)))
  280. write_excel_csv(season2009_df,"season2009.csv")
  281.  
  282.  
  283.  
  284. #holtwinters for 2007
  285. ?HoltWinters
  286. library(ggplot2)
  287. ?ts
  288.  
  289. householdtimeseries_2007 <- ts(energy_combined, frequency = 12, start = c("2007-08-01"), end = c("2007-09-01"))
  290. householdtimeseries_2007
  291. plot(householdtimeseries_2007)
  292.  
  293. august2007 <- ts(energy_combined, frequency = 1, start = c(2007,8), end = c(2007,9))
  294. august2007
  295. energy_combined
  296. View(energy_combined)
  297. View(august2007)
  298.  
  299. smooth2007<-HoltWinters(householdtimeseries_2007, beta = FALSE, gamma = FALSE)
  300. lines(fitted(smooth2007)[,1], col = 3)
  301. plot(smooth2007)
  302. plot(fitted(smooth2007))
  303.  
  304.  
  305. #Holtwinter for 2008
  306. smmoth2008<-HoltWinters(householdtimeseries_2008, beta = FALSE, gamma = FALSE)
  307. plot(smmoth2008)
  308.  
  309. #Holtwinter for 2009
  310. smmoth2009<-HoltWinters(householdtimeseries_2009, beta = FALSE, gamma = FALSE)
  311.  
  312.  
  313. View(Fore2008)
  314. View(DF2008_dec)
  315. View(DF2009_march)
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!