On November 19, JMP offered the Technically Speaking webinar, Building a Disease Outbreak Alert System Using Surveillance and Analysis. Presented by SAS’ Sam Edgemon, the webinar highlights how Edgemon used JMP control charts and the symptom database to identify a 2010 norovirus outbreak a month before the regional news networks were able to report on the outbreak.
Edgemon discusses how he uses principal component analysis to find regions in North Carolina where COVID-19 symptoms are more prevalent. He demonstrates how the use of more granular symptoms-level data combined with innovative statistical techniques has the potential to identify disease outbreaks faster while limiting false positives. Using data from emergency departments, ambulance services, poison control centers and social media, Sam demonstrated how unstructured textual data can be combined with a wide breadth of other symptom-level data to spot previously hidden outbreaks; how to identify potential "hot spots" earlier using statistical techniques not usually associated with biological surveillance; and how these techniques could be used to make a difference during the current pandemic, as well as the next one.
To help JMP sales team engage in more meaningful and productive post-event conversations with customers and prospects, we’ve created the following list of recommended resources to position.Any asset appearing on this list is considered relevant to the topic, application area or industry, performs well and is recent enough to share. To help distinguish between resources, we also include a short summary, a few words of commentary and a 4-star system to identify the "best of the best." If you have any questions about recommendations related to this event, please feel free to contact me.
This Technically Speaking webinar matches really well with the content of Sam's event. Andrea Coombs walks through an example based on building predictive models for peripheral arterial disease risk using data from the National Health and Nutrition Examination Surveys. She demonstrates how JMP, multivariate control charts, can be useful in looking at risk factors for developing a particular disease – an ability to associate risk factors with disease likelihood provides the opportunity to take preventative measures that can save lives. Along the way she also shows how to build traditional logistic models, create alternative models using more modern modeling approaches, and how to apply validation and model comparison techniques to determine the best model for disease risk prediction.
This well-attended, non-series webinar was created in response to the early surge in COVID in the U.S. It received great reviews and continues to be one of our strongest assets by all measures. Byron Wingerd uses Johns Hopkins University data to show to prepare and clean data with JMP; how to shape transform and ad features to an existing data set; how to build integrative maps and ways to effectively communicate your findings to others. Shows JMP in action in a very relatable, relevant way.
3 Stars Building and Comparing Predictive Models using JMP Pro, Mastering JMP Webinar
This Mastering JMP video discusses model types and techniques; the roles of test, training and validation data; and how to compare models. It’s relevant to this webinar because it shows how to build, validate and compare models to predict disease risk using disease and demographic data. using JMP Pro. Includes sections on Generalized and Penalized Logistic Regression, Bootstrap Forest, Boosted Tree and model comparison.
3 Stars Tackling Unstructured Data With Text Exploration, Technically Speaking
Unstructured data was a focus of Edgemon’s webinar. This Technically Speaking webinar discusses efficient ways to analyze unstructured data using JMP. Attendees will learn new advanced text exploration capabilities available in JMP that let you quickly extract insights from text-based data. The webinar also explores linkages between word usage and document of origin, how word clouds can help visualize word frequency in a way that makes clear the prevailing themes in any given unstructured data set and how to generate a Clustering Documents report.
Not a perfect match, but covers many topics of Sam's talk. This webinar shows how to use JMP Text Explorer to analyze unstructured text. You’ll learn how to interactively consolidate similar terms, recode terms, understand the underlying patterns in your text data and include the data in your analyses. The webinar covers identifying key words and phrases using term and phrase lists, word clouds and summary counts; reviewing any term or phrase in-context; interactively stemming; interpreting stem reports and stem exceptions; and recoding and parsing regular expressions.
