select_vars_for_models.qmd

---
title: "Select Variables"
format: html
---

## Setup

```{r}
library(tidyverse)
satp_data <- read.csv("data/satp_clean.csv")

glimpse(satp_data)
```

## Perpetrator

Here we look at the perpetrator of the violence. We will drop the categories of "Civilians" and "Non-Maoist armed group" as they have very few observations. Then we will save the data to a csv file along with the incident summary.

```{r}
# count number of observations in each perpetrator cateogry
satp_data |>
  count(perpetrator) # only 1 in Civilians, only 6 in non-Moaist armed group

# dropping civilian and non-Maoist armed group
perpetrator <- satp_data |>
  filter(perpetrator %in% c("Maoist", "Security", "Unknown")) |> 
  select(perpetrator, incident_summary)

write_csv(perpetrator, "data/perpetrator.csv")

glimpse(perpetrator)
```

## Action Type

Now we select all of the action types and save the data to a csv file along with the incident summary.

```{r}
action_type <- satp_data|>
  select(armed_assault:abduction, incident_summary)

write_csv(action_type, "data/action_type.csv")

glimpse(action_type)
```

## Target Type

Select the variables related to target type and save the data to a CSV file along with the incident summary

```{r}
target_type <- satp_data|>
  select(first_target:other_civilian, incident_summary)

write_csv(target_type, "data/target_type.csv")

glimpse(target_type)
```

## Deaths

Select the variables related to fatalities and save the data to a csv file along with the incident summary.

```{r}
deaths <- satp_data |>
  select(total_fatalities:other_armed_grp_fatalities, incident_summary) |>
  drop_na()

write_csv(deaths, "data/deaths.csv")

glimpse(deaths)
```

## Injuries

Select the variables related to injuries and save the data to a csv file along with the incident summary.

```{r}
injuries <- satp_data |>
  select(total_injuries:non_maoist_armed_group_injuries, incident_summary) |>
  drop_na()

write_csv(injuries, "data/injuries.csv")

glimpse(injuries)
```
## Property Damage

Just two variables here: whether there was property damage and value of property damage (if reported). I doubt the value of property damage will be useful in a model, but I'll include it for now. Save the data to a csv file along with the incident summary.

```{r}
property_damage <- satp_data |>
  select(property_damage, value_property_damage, incident_summary) 

write_csv(property_damage, "data/property_damage.csv")

glimpse(property_damage)
```
## Abductions

For kidnappings, just one variable: number of people abducted. Save the data to a csv file along with the incident summary.

```{r}
abductions <- satp_data |>
  select(total_abducted, incident_summary) 

write_csv(abductions, "data/abductions.csv")

glimpse(abductions)
```

## Arrests

Multiple arrest counts based on who was arrested. 

```{r}
arrests <- satp_data |>
  select(total_arrests:unknown_arrests, incident_summary)

write_csv(arrests, "data/arrests.csv")

glimpse(arrests)
```

## Surrenders

Total surrenders to unknown surrenders. Save as CSV.

```{r}
surrenders <- satp_data |>
  select(total_surrenders:unknown_surrenders, incident_summary)

write_csv(surrenders, "data/surrenders.csv")

glimpse(surrenders)
```

## Identifying Information

Save year, date and incident number in a separate file. 

```{r}
identifying_info <- satp_data |>
  select(year, date, incident_number, incident_summary)

write_csv(identifying_info, "data/identifying_info.csv")

glimpse(identifying_info)
```
## Location Information

Save location information, e.g. state - latitude. 

```{r}
location_info <- satp_data |>
  select(state:latitude, incident_summary)

write_csv(location_info, "data/location_info.csv")

glimpse(location_info)
```