Discovering Hidden Connection with Big Data

Picture source: https://www.bing.com
INDUSTRIAL ENGINEER – VOLUME 48: NUMBER 04
By Kim Patel, Sekar Sunararajan, Charalambos Marangos, and Emory W. Zimmers Jr.

Some companies are used to collecting data to develop standard metrics and reports. However, most of them do not use data mining tools to analyze the data further and discover unidentified cost drivers. Data mining can be used to discover hidden connections among variables. There are six steps to get meaningful insight from a big amount of data, with a case at a transportation provider company as an example:

1. Understand business objectives: provide overall goals, for example: recommending new production standard; identifying statistically valid correlations and controllable variables to improve production efficiencies; etc.
2. Form cross-functional team: having people from different department and expertise helps to get more points of view. In this case, the team consisted of maintenance organization managers, IT staff, a student, faculty and mentor consulting team.
3. Establish KPIs and set goals: develop specific KPIs for specific activities. In this case, KPIs such as how well specific activities met the engineering resource utilization standards was stated.
4. Create a master data file: identify data sets that contain particular information such as cost, labor and other recorded details of road maintenance activities, information such as location, name of foreman, equipment used, etc.
5. Conduct data mining analysis: import the sample data into the software to explore the data set and visualize trends. In this case, four analyses were performed: geographic location of maintenance activity, daily production volume, equipment type used and experience of employee.
6. Summarize findings and define opportunity: all the important findings such as correlation among variables are made known.