kafka-machine-learning-big-data

The projects predicts the human movemenrt/Postures using the live sensor data from the mobile phone sensors. This project had three phases starting with generating sensor data from the Accelerometer and Gyroscope of our phones and labeling this data.This was to be done by creating a basic Android or IOS app that collects the required data via these sensors and upload it to our Laptop via Flask Api or Bluetooth.After that this labeled data was to be stored in a database whether that be MongoDB or MYSQL.The next phase included ingesting the data in to the Kafka Environment and then processing and classifying the Labels by using ML Classification model.The last stage included implementing a frontend website using flask or any-other framework which would take Live data from our Phone and predict the position or state of the Phone and later show the readings of Accelerometer and Gyroscope on the Website along with the Predicted State. For the collection of data we developed an app using flutter and some opensource flutter api code for the implementation(refrences given at the end).The app records the X,Y and Z coordinates of both gyroscopeand accelerometer.The app takes a server address from the user and hits the data onto that server , meanwhile a Flask api has been implemented that takes the lives data being hit on that server as both the laptop and the flutter application are on the same network (in our case samew wifi connection). This is is processed , group name is added and the result is stored is a json file where the parameters of both the sensors are in a json string list.The user just had to hardcode the file name which will later be used as a label. Once we had the data in json files we had the job to upload the data to our choosen database that in our case was MongoDB. We used the concept of json file parsing as taught to us during the semester to parse the json file extract our required information and then post in on the database along with its assigned label.