Boulder Statistics assists clients from numerous industries. Over the years I have worked on projects with clients in coatings, consumer products, environmental consulting, food, marketing, medical device, medical diagnostics, pharmaceuticals, transportation, web analytics, and web usability. Below you will find a small sampling of project examples highlighting my use of graphics (produced using JMP®) in the analysis of data.
|
| Medical Device/Diagnostics |
Studies that fall under the realm of the FDA must be carefully planned,
executed, analyzed, and documented. Such studies may require sample
size determinations, statistical analysis plans, statistical sections
in an FDA submission document, as well as statistical analysis of the
study data.
Exploratory analysis is part of determining a diagnostic cut-off value that is both statistically sound and clinically meaningful. |
 |
| Exploratory Analysis for a Diagnostic Test | | |
|
|
Statistical methods can be used to analyze data from production processes for process validation, process optimization, process improvement, and process control (SPC). Projects for production sites often begin with the review of historical data, then progress to purposeful data collection to answer specific questions, and finally to long term data collection schemes for monitoring the process.
One of my favorite examples of the power of simple graphics comes from manufacturing data. It was found by looking at two histograms (shown to the left) that the highest process yields did not occur when a particular portion of the process was running at full tilt. Rather the process yield was maximized when the process speed was set at about 27 units. |
|
|
Statistical methods can be used to enhance marketing studies. Such studies might focus on consumer preferences (conjoint analysis), economic impact of events, market feasibility of a product, or the effectiveness of a marketing campaign.
In the graphic below a footer survey on a website collected "NetPromoter" scores and the shift in scores could be the result of a new product release or a site re-design. |
 |
| Average Toxin by Location | | |
Environmental studies are often data intensive. Multiple comparison techniques, such as the Analysis of Means, are often used to compare data from multiple sites under study.
In this example fish samples from three locations (A, B, and C) along a stream were taken and tested for levels of various metals. The ANOM decision chart for one particular toxin shows that fish from Site A have significantly high levels of the toxin while those from Site B have low levels when compared to the overall level of all tested fish.
|
|
