visualize.Rmd

---
title: "Garmin Data Exploration"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r setwd}
setwd("C:/Users/steve.a.cooper/Google Drive/Coursera - DS/self/garmin")
getwd()
```

## Why?

Garmin is great, but the visuals and data exploration features are pretty limited.  This script is designed to help with the upload, formating and analysis of your activity tracker... 

## Importing data

Login to Garmin connect and grab the activities in csv format.  Recommended to grab all the available fields.  Save these to your desktop and move to your working directory with the name "Activities_Master.csv"

```{r}
activities <- read.csv("Activities_Master.csv", 
                       sep = ",", 
                       skip = 2,
                       as.is = TRUE
                       )
```


## Format various fields and create a dataset for analysis

We're going to do a chunk of work to remove fields that are typically blank for running.  Sorry, I've had to hardcode a load of this for the moment

```{r}
# remove unnecessary columns
actvalues <- activities[,c(2:15, 21:22, 25:26, 32)]

# filter for running and trail running only (dplyr can only be used without POSIX fields)
library(dplyr)
actvalues <- filter(actvalues, Activity.Type == "Running" | Activity.Type == "Trail Running")

# format Elevation Gain as integer
actvalues$Elevation.Gain <- as.integer(actvalues$Elevation.Gain)

# group elevation gain into categorical variables "high" and "low"
actvalues$gain <- ifelse(actvalues$Elevation.Gain > 250, "high", "low")

# set date format for "start" field
actvalues$date <- as.Date(actvalues$Start, "%a, %b %e, %Y %H:%M %p")

# set time formats for "pace" fields
actvalues$Time <- strptime(actvalues$Time, "%H:%M:%S")
actvalues$Avg.Speed.Avg.Pace. <- strptime(actvalues$Avg.Speed.Avg.Pace., "%M:%S")
actvalues$Max.Speed.Best.Pace. <- strptime(actvalues$Max.Speed.Best.Pace., "%M:%S")


# Convert Avg.Run.Cadence to strides per minute
actvalues$spm <- as.integer(actvalues$Avg.Run.Cadence)
actvalues$spm <- actvalues$spm * 2
```

## Relationship of Cadence and Pace

Cadence has a direct relationship on pace.

```{r pressure, echo=FALSE}
library(ggplot2)
ggplot(data = actvalues, mapping = aes(x = spm, y = Avg.Speed.Avg.Pace.)) +
        geom_point() +
        geom_smooth(aes(linetype = Activity.Type)) +
        labs(title = "Runs in 2016") + labs(x = "Cadence (Strides per Minute)", y = "Pace (/km)") +
        scale_y_datetime(date_labels = "%M:%S")
```

## Evolution of Performance

```{r time_series}
library(ggplot2)
ggplot(data = actvalues, mapping = aes(x = date, y = Avg.Speed.Avg.Pace.)) +
        geom_point() +
        geom_smooth(aes(linetype = Activity.Type))
        #scale_x_datetime(date_labels = "%M:%S")
```

## Other Graphs

```{r vo2 max}

actvalues$Avg.HR <- as.numeric(actvalues$Avg.HR)

boxplot(spm~date, data=actvalues)
plot(actvalues$Avg.HR~actvalues$Elevation.Loss)

cor(actvalues)
str(actvalues)

```