Scribd Logo
Upload

Welcome to Scribd!

  • 63 views

    Sales Forecast: October 9, 2020

    Uploaded by

    salnasu
    - The document provides a sales forecast for October 2020 based on analysis of historical sales data from 1995 to 2009. - Time series models like naive forecasting and linear regression were used to predict future sales trends. - Key metrics like RMSE were calculated to evaluate forecast accuracy and monthly seasonal factors were identified.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Sales Forecast: October 9, 2020

    Uploaded by

    salnasu
    63 views9 pages
    - The document provides a sales forecast for October 2020 based on analysis of historical sales data from 1995 to 2009. - Time series models like naive forecasting and linear regression were used to predict future sales trends. - Key metrics like RMSE were calculated to evaluate forecast accuracy and monthly seasonal factors were identified.

    Original Description:

    Sales forecast in R

    Original Title

    Sales Forecast

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    - The document provides a sales forecast for October 2020 based on analysis of historical sales data from 1995 to 2009. - Time series models like naive forecasting and linear regression were used to predict future sales trends. - Key metrics like RMSE were calculated to evaluate forecast accuracy and monthly seasonal factors were identified.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    63 views9 pages

    Sales Forecast: October 9, 2020

    Uploaded by

    salnasu
    - The document provides a sales forecast for October 2020 based on analysis of historical sales data from 1995 to 2009. - Time series models like naive forecasting and linear regression were used to predict future sales trends. - Key metrics like RMSE were calculated to evaluate forecast accuracy and monthly seasonal factors were identified.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 9

    Sales Forecast

    October 9, 2020

    In [16]: #```{r results='hide', message=FALSE, warning=FALSE}


    # install all Packages required
    # install.packages('ggplot2')
    # Load Libraries to be used
    library(ggplot2)
    library(knitr)
    library(dplyr)
    library(tidyverse)
    library(summarytools)
    library(visdat)
    library(igraph)
    #library(treemap)
    #library(dplyr)
    library(devtools)
    library("readxl")
    library(repr)
    library(forecast)
    #install.packages("readxl")
    #install.packages("lubridate")
    #install.packages("forecast")
    # my_data <- read_excel(file.choose())
    #```# Do not show warnings

    Registered S3 method overwritten by 'quantmod':


    method from
    as.zoo.data.frame zoo

    In [3]: # Change plot size to 12 x 12


    options(repr.plot.width=12, repr.plot.height=12)

    In [4]: sales_df <- read_excel("Data/adfc-03.xlsx", sheet="Data")


    head(sales_df)

    1
    Date Sales
    <dttm> <dbl>
    1995-01-01 1463.9
    1995-02-01 1373.3
    A tibble: 6 Π2
    1995-03-01 1533.7
    1995-04-01 1420.6
    1995-05-01 1576.3
    1995-06-01 1607.5

    In [5]: ggplot(sales_df, aes(Date,Sales))+geom_line()

    In [12]: #create an index of time


    sales_df$time<- seq(1,nrow(sales_df))
    sales_df$month <- as.character(lubridate::month(sales_df$Date))

    2
    In [13]: monthly <- sales_df %>% group_by(month) %>% summarize(mean=mean(Sales))
    sales_df$year <- lubridate::year(sales_df$Date)
    lastYear <- sales_df %>% filter(year==2009)

    `summarise()` ungrouping output (override with `.groups` argument)

    In [14]: head(sales_df)

    Date Sales time month year


    <dttm> <dbl> <int> <chr> <dbl>
    1995-01-01 1463.9 1 1 1995
    1995-02-01 1373.3 2 2 1995
    A tibble: 6 Π5
    1995-03-01 1533.7 3 3 1995
    1995-04-01 1420.6 4 4 1995
    1995-05-01 1576.3 5 5 1995
    1995-06-01 1607.5 6 6 1995

    In [15]: install.packages("forecast")

    Installing package into 'C:/Users/woosh/Documents/R/win-library/4.0'


    (as 'lib' is unspecified)

    also installing the dependencies 'xts', 'TTR', 'quadprog', 'quantmod', 'fracdiff', 'tseries', '

    package 'xts' successfully unpacked and MD5 sums checked


    package 'TTR' successfully unpacked and MD5 sums checked
    package 'quadprog' successfully unpacked and MD5 sums checked
    package 'quantmod' successfully unpacked and MD5 sums checked
    package 'fracdiff' successfully unpacked and MD5 sums checked
    package 'tseries' successfully unpacked and MD5 sums checked
    package 'urca' successfully unpacked and MD5 sums checked
    package 'RcppArmadillo' successfully unpacked and MD5 sums checked
    package 'forecast' successfully unpacked and MD5 sums checked

    The downloaded binary packages are in


    C:\Users\woosh\AppData\Local\Temp\Rtmp6ppStO\downloaded_packages

    In [17]: sales_df.ts <- ts(sales_df$Sales,frequency = 12)

    In [21]: sales_df.ts

    3
    Jan Feb Mar Apr May Jun Jul Aug Sep O
    1 1463.9 1373.3 1533.7 1420.6 1576.3 1607.5 1545.7 1550.8 1583.0 16
    2 1580.7 1485.5 1582.6 1498.9 1644.5 1610.9 1695.0 1664.9 1583.8 17
    3 1686.0 1576.9 1608.8 1610.8 1712.9 1664.3 1698.0 1662.9 1678.9 18
    4 1724.9 1531.3 1660.0 1602.0 1677.0 1668.8 1699.6 1649.9 1685.0 17
    5 1775.4 1597.2 1733.6 1681.0 1764.5 1779.0 1817.4 1830.7 1894.3 20
    6 1924.7 1813.6 1919.9 1769.8 1985.7 2346.3 1875.7 2013.1 2015.7 20
    7 2095.8 1887.4 2063.2 1982.7 2118.3 2205.1 2217.5 2253.1 2188.7 24
    A Time Series: 15 Π12
    8 2438.8 2123.9 2296.4 2336.0 2478.6 2475.0 2465.3 2501.2 2460.6 26
    9 2528.5 2231.8 2430.7 2368.9 2539.9 2595.6 2675.1 2665.2 2726.7 29
    10 2829.9 2545.4 2746.3 2569.2 2675.2 2793.8 2843.3 2798.1 2848.0 30
    11 2963.1 2622.9 2782.9 2787.4 2879.3 2969.3 2964.4 2984.4 2971.0 30
    12 3036.0 2692.5 2953.4 2804.1 3026.2 3148.2 3120.4 3150.1 3196.1 33
    13 3332.3 2963.4 3283.3 3010.1 3239.8 3399.0 3333.5 3433.6 3443.2 35
    14 3535.1 3142.9 3212.5 3238.4 3349.2 3443.7 3413.8 3411.0 3327.8 35
    15 3560.7 2968.2 3301.6 3203.9 3379.7 3620.6 3452.3 3467.7 3406.1 36

    In [23]: plot(sales_df.ts, xlab='Month', ylab='Sales')

    4
    In [24]: ntrain = 168
    ntest = 12

    In [26]: train.ts <-window(sales_df.ts, start=c(1,1), end=c(1,ntrain))


    test.ts <- window(sales_df.ts, start=c(1,ntrain+1), end=c(1,ntrain+ntest))

    In [31]: naive.pred <- naive(train.ts, h=ntest)


    snaive.pred <- snaive(train.ts, h=ntest)

    In [33]: devtools::install_github("mlr-org/measures")

    Downloading GitHub repo mlr-org/measures@HEAD

    5
    v checking for file 'C:\Users\woosh\AppData\Local\Temp\Rtmp6ppStO\remotes41681cad3a2a\pat-s-me
    - preparing 'measures': (548ms)
    v checking DESCRIPTION meta-information
    - checking for LF line-endings in source and make files and shell scripts (543ms)
    - checking for empty or unneeded directories
    - building 'measures_0.2.tar.gz'

    Installing package into 'C:/Users/woosh/Documents/R/win-library/4.0'


    (as 'lib' is unspecified)

    In [35]: fit <- measures::RMSE(test.ts,naive.pred$mean)


    fit

    1426.6870075458

    In [36]: mod <- lm(Sales ~ time, data=sales_df)


    print(summary(mod), digits=2)

    Call:
    lm(formula = Sales ~ time, data = sales_df)

    Residuals:
    Min 1Q Median 3Q Max
    -623 -196 -48 81 1344

    Coefficients:
    Estimate Std. Error t value Pr(>|t|)
    (Intercept) 1299.41 46.40 28 <2e-16 ***
    time 13.48 0.44 30 <2e-16 ***
    ---
    Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    Residual standard error: 310 on 178 degrees of freedom


    Multiple R-squared: 0.84, Adjusted R-squared: 0.84
    F-statistic: 9.2e+02 on 1 and 178 DF, p-value: <2e-16

    In [38]: pred <- predict(mod,sales_df)


    plot(sales_df$Sales, ylab='Sales', xlab='Time')
    lines(pred, lwd=1,col="blue", lty=1)

    6
    In [40]: mod <- lm(Sales ~ time + I(time^2) + factor(month), data=sales_df)
    print(summary(mod), digits=2)

    Call:
    lm(formula = Sales ~ time + I(timeˆ2) + factor(month), data = sales_df)

    Residuals:
    Min 1Q Median 3Q Max
    -441 -79 6 98 464

    Coefficients:
    Estimate Std. Error t value Pr(>|t|)

    7
    (Intercept) 1.4e+03 4.7e+01 30.2 <2e-16 ***
    time 9.3e+00 8.2e-01 11.4 <2e-16 ***
    I(timeˆ2) 2.2e-02 4.4e-03 5.0 2e-06 ***
    factor(month)10 6.9e+01 5.2e+01 1.3 0.188
    factor(month)11 1.5e+02 5.2e+01 2.8 0.005 **
    factor(month)12 8.4e+02 5.2e+01 16.2 <2e-16 ***
    factor(month)2 -2.7e+02 5.2e+01 -5.3 4e-07 ***
    factor(month)3 -1.2e+02 5.2e+01 -2.3 0.025 *
    factor(month)4 -2.1e+02 5.2e+01 -4.1 7e-05 ***
    factor(month)5 -8.1e+01 5.2e+01 -1.6 0.121
    factor(month)6 -8.8e+00 5.2e+01 -0.2 0.865
    factor(month)7 -5.6e+01 5.2e+01 -1.1 0.282
    factor(month)8 -5.5e+01 5.2e+01 -1.1 0.293
    factor(month)9 -7.0e+01 5.2e+01 -1.3 0.180
    ---
    Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    Residual standard error: 140 on 166 degrees of freedom


    Multiple R-squared: 0.97, Adjusted R-squared: 0.97
    F-statistic: 3.9e+02 on 13 and 166 DF, p-value: <2e-16

    In [42]: pred <- predict(mod,sales_df)


    plot(sales_df$Sales, ylab='Sales', xlab='Time')
    lines(pred, lwd=1,col="blue", lty=1)

    8
    In [ ]:

    You might also like