[1] A. Reiss and D. Stricker. Introducing a New Benchmarked Dataset for Activity Monitoring. The 16th IEEE International Symposium on Wearable Computers (ISWC), 2012.
Install all the necessary programs and tools (these instructions are for homebrew. If you use apt-get, that also works).
brew install pip
pip install flask
Clone the git repository:
git clone https://github.com/alexchow/PAMAP.git
cd PAMAP
Download the data set:
wget http://archive.ics.uci.edu/ml/machine-learning-databases/00231/PAMAP2_Dataset.zip
unzip PAMAP2_Dataset.zip
Populate the sqlite database with the dataset
python start.py populatedb
Note: if the python process is getting killed, it's probably linux's Out-of-memory killer. Set up a swap file
Run the server:
python start.py
You can see data and features visualized in your web browser at
http://localhost:5000/view
In the source code, you can see that it uses the URL for deciding the page view. Queries accepted are as follows:
Query | Value |
---|---|
windowInterval | The number of samples in each time window. Default = 200 |
activities | Comma separated list of activity IDs to use. Default = 'handAccX', handAccY,chestAccX,chestAccY,ankleGyrX,heartrate |
data_keys | Comma-separated list of raw data columns to use. Example: handAccX,chestGyrY,heartrate. Default uses all sensor data |
numWindows | The number of time windows to use. Default = 3 |
An example query is:
http://localhost:5000/view?windowInterval=250&numWindows=4&data_keys=handAccX,chestAccX,ankleGyrX