processMining-v02.Rmd

---
title: Process Mining
output: 
  rmdformats::readthedown:
    toc_depth: 4
    number_sections: TRUE
    self_contained: true
---

```{r setup, include=FALSE}

knitr::opts_chunk$set(echo = FALSE, comment = FALSE, message = FALSE, warning = FALSE, error = FALSE, fig.width = 9, fig.align = "center")

options(knitr.table.format = "html")


library(dplyr)
library(lubridate)
library(gt)
library(ggplot2)
library(bupaR)
library(daqapo)
library(processmapR)

```

```{r readData&Format}

#read data
data <- read.csv("./IncidentData.csv")

#clean header
data <- data %>% janitor::clean_names()

#data date format
data$date_stamp <- ymd_hms(data$date_stamp)

# bupar requirement to create life status column

data$lifecycle <- "Start"

# merge incident value
data <- data %>% distinct() %>% arrange(date_stamp) 

# activity instance id created
data$act_instance_id <- 1:nrow(data)


```

# Data Summary  

```{r}

start_date <- min(as.Date(data$date_stamp))
end_date <- max(as.Date(data$date_stamp))

data %>%
  head(10) %>% 
  DT::datatable()

```

Number of tickets to analyze **`r length(unique(data$incident_id)) `** with **`r ncol(data)`** attributes And total number of events **`r nrow(data) `**

## State Occurance

> Top 10 State

```{r}

data %>% 
  count(incident_activity_type) %>% 
  top_n(10) %>% 
  ggplot()+
  geom_bar(aes(x=n,y=reorder(incident_activity_type,n)),stat="identity")+
  labs(x="Count",
    y ="Activity type")+
  theme_classic()

```

> Bottom 10 State

```{r}

data %>% 
  count(incident_activity_type) %>% 
  top_n(-10) %>% 
  ggplot()+
  geom_bar(aes(x=n,y=reorder(incident_activity_type,-n)),stat="identity") +
  labs(x="Count",
    y ="Activity type")+
  theme_classic()



```   

## Most No. of Transition

```{r}

data %>% 
  count(incident_id) %>% 
  top_n(20) %>% 
  ggplot()+
  geom_bar(aes(x=n,y=reorder(incident_id,n)),stat="identity") +
  labs(x="Count",
    y ="Incident id")+
  theme_classic()

```

```{r eventlog}

# change colname for date
colnames(data)[2] <- "start"

event_log <- eventlog(data,
                      case_id = "incident_id", 
                      activity_id = "incident_activity_type",
                      activity_instance_id = "act_instance_id",
                      lifecycle_id = "lifecycle",
                      timestamp = "start",
                      resource_id = "assignment_group")

```

# Data Quality Check

```{r activityLog}

activity_log <- daqapo::activitylog(activitylog = data,
                               case_id = "incident_id", 
                               activity_id = "incident_activity_type",
                               lifecycle_ids = c("start"),
                               resource_id = "assignment_group")

```

> Missing Value 

```{r}

activity_log %>% 
  detect_missing_values()

```

> Duration Outlier

```{r}
# 
# activity_log %>% 
#   detect_inactive_periods(threshold = 30)

```

> Overlaps

```{r}
# 
# activity_log %>% 
#   detect_overlaps()

```
> Time Anomalies

```{r}

# activity_log %>% 
#   detect_time_anomalies()

```

# Event Data

## Event Mapping 

```{r}

mapping_fines <- mapping(eventlog = event_log)
mapping_fines

```

## Event Summary 

```{r}

event_log %>% 
 summary()

```

## Trace Detail

```{r}

event_log %>% 
traces %>% 
  DT::datatable()


```

## Trace Map

> Top 20 trace

```{r}

event_log %>% 
  trace_explorer(n_traces = 10,type = c("frequent"))

```

 > Bottom 20 trace

```{r}

event_log %>% 
  trace_explorer(type = c("infrequent"),n_traces = 10)  

```

# Process Map

## Frequency Profile

> Relative Frequency

* The relative number of instances per activity
* The relative outgoing flows for each activity

```{r}

event_log %>% 
filter_trace(c(1:10)) %>% 
 process_map(type = frequency("relative"))

```

> Relative Case

* The relative number of cases per activity and flow

```{r}

event_log %>% 
filter_trace(c(1:10)) %>% 
 process_map(type = frequency("relative_case"))

```

> Absolute Frequency

* The absolute number of activity instances and flows

```{r}

event_log %>% 
filter_trace(c(1:10)) %>% 
 process_map(type = frequency("absolute"))

```

> Absolute Case Frequency 

* The absolute number of cases behind each activity and flow

```{r}

event_log %>% 
filter_trace(c(1:10)) %>% 
 process_map(type = frequency("absolute_case"))

```

## Performance Profile

>Mean Profile


```{r}

event_log %>% 
  filter_trace(c(1:10)) %>% 
   process_map(performance(mean, "hours"))

```

## Combination Profile

```{r}

event_log %>% 
  filter_trace(c(1:10)) %>% 
    process_map(type_nodes = frequency("relative_case"),
                type_edges = performance(mean))

```

# Resource Map

```{r}


event_log %>% 
  filter_trace(c(1:10)) %>% 
  resource_map()

```


# Precedence Matrix

>Absolute frequency

```{r}


event_log %>%
  filter_trace(c(1:10)) %>% 
    precedence_matrix(type = "absolute") %>%
    plot

```

>Relative Frequency

```{r}
event_log %>%
  filter_trace(c(1:10)) %>% 
    precedence_matrix(type = "relative") %>%
    plot



```

>Antecedent-wise Frequencies

```{r}
event_log %>%
  filter_trace(c(1:10)) %>% 
    precedence_matrix(type = "relative-antecedent") %>%
    plot



```

>Consequent-wise Frequencies

```{r}

event_log %>%
  filter_trace(c(1:10)) %>% 
    precedence_matrix(type = "relative-consequent") %>%
    plot

```