Inspiration
We joined the UPM challenge because we want to improve sustainable development by making UPM's processes more efficient and we were keen on developing our data analytics skills. .
What it does
Our work analyzes the UPM data and gives several example insights that can be gleaned from the data including in both production and customer claims areas.
How we built it
We utilized several different data analysis tools including R, Excel, and Spark (on IBM bluemix).
Challenges we ran into
A lack of revenue data makes it hard to predict the actual improvement and profitability but we have tried to provide a guess estimate. Since we weren’t familiar with paper industry, understanding the data and mill processes took a significant part of time which would otherwise be spent on data analytics.
Accomplishments that we’re proud of
We were able to improve the mean error in prediction of unplanned break durations from a baseline by 20%. We hope that such data analysis that spans over all UPM mills will improve overall planning efficiency.
What we learned
We learned about the paper production process and the importance of having good contextual information (such as the information provided by the onsite UPM experts) that can help guide the data analysis. In other words, data and big data analysis is relatively useless without clear and detailed domain knowledge.
What's next for IOT_UPM_Chall3_ParadigmSisters
We are all part of a university research team and thus we will continue to collaborate on data analysis in the university. We hope to compete together again in future hackathons to improve further!
Log in or sign up for Devpost to join the conversation.