Coffee Roaster Script

05 May 2019

Silaa Farms is a socially responsible company located in North Thailand that partners with smallholder farmers to produce premium quality organic Arabica coffee.
  • The need: take raw CSV data hosted on a LAN device, divide a day's data into roasts, and convert the data for use in Artisan
  • The solution: use beautiful soup to read changing filenames off the LAN device
  • Make a small UI in TKlinter to give a worker information on whether running the script worked (it has to be manually run once a week)
  • use a python library to pull any new files
  • Write a script to cut the data
  • Write a script to convert the data
  • Use a python - google API to upload the finished files (If they are new) to a google drive account
  • Result: Use can access individual roast data from anywhere in the world

The core values of Silaa Farms are quality and consistency. They have made it a priority to build quality control checks into every stage of their production process, and have trained their employees to understand and thoroughly perform each check.
Maintaining consistency in the coffee roasting stage was one quality control area Silaa Coffee wanted to improve. Located in a backwater village in northern Thailand, the 500kg roaster collects sensor and spin data while roasting, and stores it locally. This data can be downloaded directly by accessing the IP of the roaster and viewing an HTTP interface. Currently, an expert visits the village every week or so and downloads the data from the coffee roaster onto a laptop and then takes it away to analyze. However, visiting so frequently was becoming burdensome, and they wanted to have a local employee be able to pull the data from the roaster and put it online for the expert to analyze.

The goal for us was to create a simple, hopefully, automated solution--ideally one which was very easy and simple to use--to upload data about all the roasts from the coffee roaster to Google Drive so they can be quickly looked at for reference and quality control. This challenge was dubbed #InternetOfBeans.
During the Hackathon, our team tackled this problem. Rather than requiring a laptop and heavy manual processing, the approach taken was to invest in a Raspberry Pi and give it a button and LED. The roaster presses the button just before it turns itself off. The LED lights up for at least two seconds, during which the data is pulled from the roaster and uploaded to the internet. Once the data is safely transferred, the LED is extinguished and the roaster can turn off the machine. This way the data is synced to Google Drive and can be accessed remotely anytime.
Now, the software component of the project has been implemented and is working well. It is saving roast information that was getting lost after a certain number of days, dividing up the day's data into individual roasts, and converting that information into usable data. With the data being synced to Google drive automatically, it can be accessed remotely with one click and easily analyzed to get a good feel for the consistency of the roasts. This all ensures that the coffee continues to be the high-quality product that Silaa Farms is known for.