Inspiration
Be able to make sense of large amount of timeseries data coming from sensors of various components installed in smart home system
What it does
Visualize the cause of failures in the components
How we built it
We understood the product which is heat pump by Wolf. We started with understanding of data fields in the time-series data. There were overall 3 types of files and this needed to be merged in order to find linked data to predict the cause of failure in a heat pump. We used python to perform data analysis. On a parallel track we also thought of showing the analysed data through a web page with graphs plots. We looked at some technologies and thought of using React for the same.
Challenges we ran into
Data field mappings Data is not cyclic and asynchronous (main hurdle in combining data from the files) Data field representations New Technologies (GraphQL, React, OAuth, etc.) Time-series analysis
Accomplishments that we're proud of
Understanding of the working system architecture Analysis of time series data
- Finding error timing and the devices linked to those errors
What we learned
New Technologies Analysis of Time-series data
What's next for Failure Detection
Spend more time on cleaning the data and removing non-essential noises Data Visualization in a user friendly way
Log in or sign up for Devpost to join the conversation.