Case studies
01
Transforming Cloud Data Into Usable Models and Visualizations
We worked with the manufacturer of an internet-connected consumer device who asked to set up a pipeline to clean and transform Amazon cloud data from individual devices. We also did exploratory data analysis, visualization, and modeling to understand consumer usage and repeat patterns. Using Python scripts, we developed a consumption measure and modeled consumption by device. These allowed us to create visualizations to serve as a prototype dashboard for their team, leadership, and suppliers. Ultimately, we used creative application of regression and statistical modeling to model phases of consumer familiarity with the product and to create models for consumption based on device settings and usage.
Services Provided
AWS data transformation and visualizationExploratory data analysisStatistical modelingPowerBI Dashboard creation and ongoing updates
02
Uncovering Sales Patterns Through Data Visualization
A medical supply company asked for help uncovering sales patterns in their customer relationship (CRM) data across sales channels, customers, and a 1000+ item product catalog. We developed scripts and manufacturing-inspired visualizations to create measures for repurchase of frequently-sold items. This gave them valuable insight into customers’ purchasing habits. With this new information, the company was able to focus on key customer/product combinations and direct specific sales attention to areas that were already working well and could be expanded and honed to generate new business.
Services Provided
CRM data analysis pipelineData visualizationStatistical repurchase measures
03
Creating a Flexible Software Architecture for a Growing Business Segment
A company’s computer Modeling and Simulation R&D group asked for help with redesigning the user interface, data structure, and code base for an important application that is the user-facing front end for a virtual simulator used for both understanding current products and developing new formulations. We used their existing system and its extensive code as the jumping off point for getting this complex application to a usable and flexible form. This ongoing project will let them explore new product formulation and processing space related to upstream product designs. Under the hood, we created an object-oriented, maintainable code base that can be expanded further as their needs grow.
04
Troubleshooting and Business Opportunity Development Through Data-Rooted Problem Definition
Troubleshooting and business opportunity development often need to be more complex than simply pulling a known lever to reset a problem back to baseline. Business opportunity development is always more open-ended than that. We use a five-step methodology to solve unexpected in-market, manufacturing and/or material supply crises. This approach also works to go after complex new opportunities. We start with a business problem such as “the line suddenly started having very high scrap” or “This market segment started losing volume.” We create a multi-faceted problem statement that is engaging to diverse experts and which leads to brainstorming and a learning plan. Our methodology engages leadership, keeps them informed and gives experts space to do the work while remaining collaborative with leadership. Our approach has led to breakthroughs in consumer complaints and in manufacturing quality crises where the solution needed to be a combination of business practice, product design, raw material, and/or operational changes.
Services Provided
Facilitated brainstormingProblem statement creationData-rooted change planning
05
Teaching Data Technique Fundamentals to Enable Digital Transformation
We developed a Data Techniques Fundamentals training to bring about grass roots digital transformation in organizations. This is flexible but has been offered as a combination of large group training followed by consultation with individual project teams to tailor to their needs. Through trainings like this, we provide engineers, scientists, and their leaders with the fundamentals to work efficiently and collaboratively with all kinds of experimental data in an R&D culture or to build exploratory data skills in a manufacturing culture. Hands-on classes can focus on a toolset of Excel, JMP, PowerBI and/or Python/Pandas.. We share a standard workflow for gathering and curating raw data thereby increasing the collaborative, useful life of data. Based on experience instructing groups of up to 80 attendees, this training also gives managers and leaders a compass for implementing repeatable, collaborative systems in a learning organization.