Week12_DD

library(readxl)

library(readr)

library(dplyr)

library(lubridate)

library(tsibble)

library(ggplot2)

library(fable)

library(feasts)

library(slider)

Load the dataset

ufc_data <- read_excel(“~/Downloads/UFC_Dataset.xls”)

ufc_data_subset <- ufc_data |>

select(Date, RedAvgSigStrLanded, BlueAvgSigStrLanded) |>

mutate(

Date = case_when(

!is.na(mdy(Date)) ~ mdy(Date),

!is.na(dmy(Date)) ~ dmy(Date),

!is.na(ymd(Date)) ~ ymd(Date),

TRUE ~ as.Date(NA) # If parsing fails, set to NA

),

AvgStrikesLanded = rowMeans(cbind(RedAvgSigStrLanded, BlueAvgSigStrLanded), na.rm = TRUE)

) |>

# Remove rows with NA dates or missing values for analysis

filter(!is.na(Date), !is.na(AvgStrikesLanded)) |>

group_by(Date) |>

summarize(AvgStrikesLanded = mean(AvgStrikesLanded, na.rm = TRUE))

Check for duplicates

ufc_data_subset <- ufc_data_subset |>

distinct(Date, .keep_all = TRUE)

Ensure the Date column is properly formatted

ufc_data_subset <- ufc_data_subset |>

mutate(Date = as_date(Date))

Create tsibble

ufc_tsibble <- ufc_data_subset |>

as_tsibble(index = Date)

Check the tsibble

print(ufc_tsibble)

Plot the average strikes landed over time

ggplot(ufc_tsibble, aes(x = Date, y = AvgStrikesLanded)) +

geom_line() +

labs(title = “Average Significant Strikes Landed Over Time”,

x = “Date”, y = “Average Strikes Landed”) +

theme_minimal()

Linear regression for strikes over time

linear_model <- ufc_tsibble |>

model(trend_model = TSLM(AvgStrikesLanded ~ trend()))

report(linear_model)

linear_fc <- linear_model |>

forecast(h = “1 year”)

autoplot(ufc_tsibble, AvgStrikesLanded) +

autolayer(linear_fc, series = “Trend Forecast”) +

labs(title = “Linear Trend of Average Strikes Landed”,

x = “Date”, y = “Average Strikes Landed”) +

theme_minimal()

Smoothing to detect seasonality

ufc_tsibble |>

mutate(MA_AvgStrikesLanded = slide_dbl(AvgStrikesLanded, mean, .before = 4, .complete = TRUE)) |>

ggplot(aes(x = Date)) +

geom_line(aes(y = AvgStrikesLanded), color = “blue”, alpha = 0.5) +

geom_line(aes(y = MA_AvgStrikesLanded), color = “red”) +

labs(title = “Average Strikes Landed with Smoothing (Moving Average)”,

x = “Date”, y = “Average Strikes Landed”) +

theme_minimal()

Seasonality detection using ACF/PACF

ufc_tsibble |>

ACF(AvgStrikesLanded) |>

autoplot() +

labs(title = “ACF for Average Strikes Landed”)

ufc_tsibble |>

PACF(AvgStrikesLanded) |>

autoplot() +

labs(title = “PACF for Average Strikes Landed”)

I give up, I think i was doing something wrong with the tsibble.