Vincent L. Cannataro

ECfitbitR

You can use the ECfitbitR package to access your FitBit data through the FitBit API and format your data for easy analysis. This package relies heavily on the excellent fitbitr package to do the heavy lifting of interacting with the API. Please visit fitbitr for more details of the extensive functionality therein, as ECfitbitR is designed with limited but specific functionality in mind.

First things first

Creating a FitBit web app

The first thing you need to do to access your own FitBit data is make a personal FitBit API web app.

1. Go to https://dev.fitbit.com/apps/new and log into your FitBit account.
2. Register your application. This is your own personal app for your own personal data, so you do not need a terms of service to share with clients. Hence, we just use the website of the R package (https://github.com/vcannataro/ECfitbitR) for our terms. IMPORTANT: the fields must exactly match what is in this image. Feel free to copy+paste my inputs:

• https://github.com/vcannataro/ECfitbitR
• http://localhost:1410/ (for Callback URL. Note that this needs to be http: and not https:)

Also make sure to set your OAuth 2.0 Application Type as Personal.

3. Retrieve your Client ID (FITBIT_KEY) and your Client Secret (FITBIT_SECRET). You can find these here: https://dev.fitbit.com/apps.

You are now ready to access your data! Well, first you need a toolkit to use. We will use the free, open source, statistical computing platform R.

Downloading R

1. Go to https://cloud.r-project.org/ and select your specific download for your specific operating system.
2. Follow the instructions to download and install R.

Installation

ECfitbitR is a mostly self-contained package, but it does have some dependencies that need to be installed.

1. Open R
2. Click File –> New Document (or whatever the equivalent is in your operating system).
3. Copy and paste the following into the document. You may be prompted to select a “mirror” server to download the packages from—choose either the top one or whatever location is closest to you (this mattered more when download speed was highly dependent on distance to servers).

install.packages( "https://cran.r-project.org/src/contrib/rlang_0.4.1.tar.gz", repos = NULL, type="source")
install.packages( "https://cran.r-project.org/src/contrib/digest_0.6.22.tar.gz", repos = NULL, type="source")
install.packages(c("stringr","lubridate","dplyr","tidyr","purrr","httr","jsonlite","httpuv","RCurl","devtools","magrittr"))

4. On the line of the code that you just copied, press CMD+return (or ctrl+R on PC, etc.)

• If prompted for a mirror (where the download will come from), click the top cloud or whatever is closest to your location
• If prompted with a message about “Do you want to install from sources the package which needs compilation? (Yes/no/cancel)” type Yes and hit enter.

5. On a new line in your document, copy the following and then repeat step 4. NOTE: If on PC, you may be asked to update a certain Rtools package, or to update existing packages. Just hit Enter to skip the update at the prompt.

devtools::install_github("teramonagi/fitbitr"); devtools::install_github("vcannataro/ECfitbitR")

You are now ready to download your data!

Every time a chunk of code is presented below in a gray box, you can either copy it directly into the R terminal and press return or Enter, or you can develop your own R script using a New Document (step 2 above) and press CMD+return or ctrl+R to send it to the terminal. If you copy multiple lines into the R script you can send the code to run in the terminal line-by-line, or you can highlight a chunk of code and then hit CMD+return. The advantage of having theR script is that you can save the script and return to it later—allowing you to see the exact steps you took to get to your end product (very important in the reproducibility of science and analyses)!

Every time you see the # sign in the code it is a “comment” and R automatically does not run comments.

As a general note, you need to run the following code chunks in order because much of the subsequent code depends on the code prior to it.

Downloading your FitBit data

Initializing the API

The first thing we will do is tell FitBit who you are. For this, we need the FITBIT_KEY and FITBIT_SECRET generated in the First things first step.

# load the key and secret into your R environment. 
FITBIT_KEY <- "your fitbit key pasted here in the quotes"
FITBIT_SECRET <- "your fitbit secret pasted here in the quotes"

Next, we will authorize ourselves to obtain our own data. After running the next block of code your internet browser will prompt you to authorize your app. Highlight relevant fields and click Allow. It will look like this:

token <- fitbitr::oauth_token()

Downloading your data from fitbit

Run the following code to download your data from fitbit:

# for the month of October, 2019:
my_data <- ECfitbitR::get_my_data(token=token, start_date = "2019-10-31")

# for the previous month from today's date (remove the `#` to run): 
# my_data <- ECfitbitR::get_my_data(token=token)

You can save all of your data to your computer in an R data file (.RData). Either click File --> Save Workspace (PC) or Workspace --> Save workspace file... (MAC) and find a directory to save your data. Or, follow the code below. Make sure to keep the my_fitbit_data.RData, or pick your own name, so that your file has a name. I am saving directly to my Desktop:

# for MAC: 
save(my_data, file = file.path("~","Desktop","my_fitbit_data.RData")) 
message("MAC users have their data saved to: ",file.path("~","Desktop"))

# for PC:
save(my_data, file = file.path(path.expand("~"),"my_fitbit_data.RData",fsep="\\")) 
message("PC users have their data saved to: ",file.path(path.expand("~")))

You can also save your data as a .csv to open in other programs, like Microsoft Excel.

# saving heartrate data summary
# For MAC: 
write.csv(x = my_data$my_heartrate_data$heartrate_daily_summary,file.path("~","Desktop","my_heartrate_data.csv"), row.names = F)

#For PC: 
write.csv(x = my_data$my_heartrate_data$heartrate_daily_summary,file.path(path.expand("~"),"my_heartrate_data.csv",fsep="\\"), row.names = F)

Clean and save your data for analyses

Much of the data in my_data is in per-minute resolution. The ECfitbitR::clean_my_data function will split this data into 6 hour blocks for you for easy analyses.

my_clean_data <- ECfitbitR::clean_my_data(my_fitbit_data = my_data)

Finally, you can save your data with the ECfitbitR::save_my_data function. This function defaults to making a my_fitbit_data directory on your Desktop.

# For MAC: 
ECfitbitR::save_my_data(raw_data = my_data, clean_data = my_clean_data,MAC_or_PC = "MAC")

# For PC: 
ECfitbitR::save_my_data(raw_data = my_data, clean_data = my_clean_data,MAC_or_PC = "PC")