The first Insight Net Technology Transfer Workshop was held on December 11 and 12, 2024, in Chapel Hill, NC. The goal of the tech transfer series is to share and provide training to public health employees on tools and capabilities created by Insight Net developers. This workshop focused on three tools developed by Delphi Group: epidatr, epiprocess, and epipredict, which are designed to help users access and interpret aggregate data, address data challenges associated with reporting delays, outliers, and data noise, create basic nowcasts and forecasts and test predictions. All workshop tools and instructional materials are public and can be found on GitHub here. Workshop presentation materials are here. The workshop curriculum was developed and taught by Delphi Co-Primary Investigator Ryan Tibshirani (UC Berkeley), Co-Investigator Daniel Macdonald (UBC), and core members Rachel Lobay (UBC), and Logan Brooks (CMU). 35 trainees attended the workshop, including 25 public health employees (epidemiologists, data scientists, and program officers) and 10 academics in public health who work with public health departments.
Workshop participants completed pre and post-workshop surveys regarding their workplace needs and skills related to the workshop. In the pre-workshop survey, completed by 20 public health attendees, respondents were asked to rate the importance of the following skills to their work: data cleaning, analysis, and visualization, forecasting and nowcasting. Additionally, they were asked to rate their own familiarity and skills in 6 related areas: data access, data processing, data visualization, data analyzing, creating forecasts, and overcoming issues with data inconsistencies. In the post-workshop survey, participants were asked again to rate their familiarity and skills in these areas (using a code that connected to their pre-survey responses). Results of the survey showed increases along three dimensions: analyzing data, creating forecasts, and overcoming issues with data inconsistencies and gaps. There was a minor decrease in score for completing data processing steps, which could be because respondents became aware of additional data processing steps they had not considered before the pre-survey. Finally, when asked on the post-workshop survey if and how they plan to use Delphi’s tools in their work, five respondents reported plans to use the package to improve data access. Nine plan to use the tools to build forecasting or nowcasting models to do things such as (1) nowcast monkeypox and dengue from prospective testing to determine the number of cases missed or asymptomatic, (2) more efficiently deploy resources, (3) validate and compare performance of nowcasting approaches their jurisdiction has developed and (4) help manage syphilis outbreaks.