A health-based recipe and exercise plan recommendation system. A Line chatbot that recommend recipes based on user's ingredient preference, daily calories consumption, recipes style preference and other users preference behavior. Users can also send a photo of gym equipment, the chatbot will tell you what the equipment is and provide a link to a website that teaches you how to exercise with that equipment.
Demo Video: 食健主義Demo Video
Demo PowerPoint: 食健主義PowerPoint
All services and databases are built in docker containers, including python devops environment, mongoDB, MySQL and kafka.
- Raw data collected from the web --> store in mongoDB.
- Push data to Hadoop file system that runs on local machines.
- Utilize SparkSQL to preprocess our datas, and Spark Mllib for model training.
- Use Tensorflow for image recognition model training.
- Build a Line Chatbot App in python devops docker container, with pipenv for libraries version control.
- Construct docker-compose.yml file to run all containers.
- Connect all containers by port mapping.
- Follow the chatbot.
- Richmenu
- Recipe recommendation.
- Gym Equipment image recognition.
- Functions as Saving workout plans or Saving recipes are also included. More demo images, please refer to demo_image directory.