Garmin Forerunner 205

Using the Garmin Forerunner 205 with R on Ubuntu 10.10

Below I describe how I use Ubuntu and the GNU R statistical computation and graphics system for analyzing data from my jogging tracks. In particular I describe how to

  • copy the data from the Garmin Forerunner to a sqlite database on my Ubuntu system
  • plot course data from the database using R
  • retrieve Google maps using R and overplot my course

In my example I use a sqlite database to store the track data, but using e.g. postgresql is just as easy. Just install the respective packages and change the configuration of your ODBC data source accordingly.

Prerequisites

  1. Ubuntu packages which need to be installed
    • garmin-forerunner-tools
    • r-base-core
    • sqlite3
    • python-sqlite
    • unixodbc-dev
    • libsqliteodbc
    • libgdal1-dev
    • libproj-dev
  2. R packages which need to be installed
    • RODBC
    • png
    • RgoogleMaps
    • sp
    • rgdal

For each of the required R packages, download the source .tar.gz file from CRAN - the comprehensive R archive network and install using

sudo R CMD INSTALL -l /usr/lib/R/site-library <package>.tar.gz

Create the sqlite3 database

$ sqlite3 jogging.db
SQLite version 3.7.2
Enter ".help" for instructions
Enter SQL statements terminated with a ";"
sqlite> CREATE TABLE rounds (
           id INTEGER PRIMARY KEY,
           date timestamp,
           round varchar(64));
sqlite> CREATE TABLE gmn (
           rid INTEGER,
           time timestamp,
           latitude float,
           longitude float,
           altitude float,
           distance float);

Adding tracks

Copy your latest track from the Forerunner to your local disk using the utility that comes with the garmin-forerunner-tools:

$ garmin_save_runs

How to add a track to the jogging database:

$ ./addround.py <description> <gmn-file>

E.g.:

$ ./addround.py Möllmicke 2011/01/20110123T093114.gmn

where the python script addround.py uses the tool garmin_dump to translate the gmn-file to ASCII and writes the parsed output to the sqlite database.

To check available track data:

$ sqlite3 jogging.db
sqlite> select * from rounds;

Make the sqlite database available in R by editing your .odbc.ini file to contain

[jogging]
Driver = /usr/lib/odbc/libsqlite3odbc.so
Setup = /usr/lib/odbc/libsqlite3odbc.so
Database = /path/to/your/jogging.db
[ODBC]
InstallDir = /usr/lib

To plot our data in R, run the script garminsql.R in your R environment:

$ R
> source("garminsql.R")
> rnd <- plotround(<id>, download=TRUE)

where is the primary key in table ''rounds'' that was generated when inserting the data into the database. The variable ''download'' controls if a google map should be downloaded and used as background for the plot of the track. By using FALSE, you can skip the download and run the script without being connected to the internet.

The resulting plot should look similar to this one:

three panel R plot

External links