Analytical tools facilitate teamwork dramatically decreasing the amount of man-hours required to achieve their objective.
A heavy industrial equipment maker and operator needed to collapse the endless cycles involved in building reliability models.
Being data driven can be very time consuming. There are challenges in acquiring, integrating and structuring data and then isolating irrelevant data and outliers. Data preparation cycles are often long and very tedious.
In one recent case, a group of engineers were tasked with building reliability models to predict maintenance and downtime on large, industrial equipment. Building these models was extremely time consuming taking them weeks to build these models. Each meeting ended with a request to put together yet another variant set of data and arrange yet another meeting. These cycles would continue, it seemed, indefinitely. Schedules were disrupted and there was a loss of history from one meeting to the next.
We built an interactive, integrated tool that pulled and massaged the data, built models, and allowed easy comparison (both graphical and with modeling metrics) to compare models. Many tools build single models – this tool builds multiple models so they can readily choose one that is the best. The tool also provides the means to create variant subsets of data so decisions can be made about their relevance and representation.
Predictum’s integrated tool allowed for rapid development of models based on interactively creating variant sets of data dramatically reducing the amount of time and effort engineers spent in modeling.
Analysis now happens with experts acting collectively, in real time. Numerous meetings and weeks of work ended up taking hours with better results. Decisions are no longer deferred time and time again. Because data can be so easily handled, sensitivity analysis on the contribution of each data point is easy with instant updates on the models.
A group of engineers required an endless number of meetings to make decisions.
An integrated, interactive tool that allows engineers to dynamically model variant sets of data in real time.
Endless series of meetings replaced with one, concise session where decisions are based on full exploration of available data.