Scripts to capture OBD-II data into JSON format for analytics.
Additional scripts and service examples are provided for JSON logging from gpsd/gpspipe (gpspipe.py) and accelerometer data (9dof.py) using Adafruit 9-DOF ACCEL/MAG/GYRO+TEMP BREAKOUT BOARD - LSM9DS1 from https://www.adafruit.com/product/3387
You'll need an OBD-II to USB serial interface to read the data. I'm using a Scantool OBDLink SX.
The scripts are developed and tested on Raspberry Pi platform. It should work with any model Pi, though you'll need a car USB charger for power.
If you'd also like to collect GPS data to complement the OBD-II readout, there's a service and script to collect data from a USB GPS device supported by gpsd. I'm using a u-Blox 7 GPS/GLONASS receiver.
Install Python 3 and the obd package (pip3 install obd)
Plug in the OBD-II connector and try running the script with "python3 obd2json.py". It should start outputting JSON data to a log file in the /home/pi directory.
To start the script automatically, copy the obdii.service file to /lib/systemd/system and run:
sudo systemctl enable obdii
You can also start it immediately with sudo systemctl start obdii
If you have a GPS device attached, you can add the included gpspipe service and script to collect GPS data. The GPS and OBD data will be logged as JSON to separate files currently.
The obd library will scan for valid codes, and the script will read all Mode 1 diagnostic codes continuously (or, as fast as the interface can supply them, plus a short delay).
The library is documented at https://python-obd.readthedocs.io/en/latest/
The JSON output can be read by any engine that supports JSON input. For example, load the data into Vertica (http://www.vertica.com/) with:
CREATE FLEX TABLE obd();
COPY obd FROM LOCAL 'obd-file' PARSER FJSONPARSER();
SELECT COMPUTE_FLEXTABLE_KEYS_AND_BUILD_VIEW('obd');
-- If using vsql, you may wish to set \x to read the OBD fields more easily
SELECT * FROM obd_view LIMIT 10;