 This elevator pitch will discuss how ours being used to improve the processes at the Public Health Agency of Canada during the COVID-19 pandemic. To effectively manage public health needs at the federal level during this pandemic, the agency needs to be meticulous from a surveillance and reporting standpoint. Surveillance needs to be performed to gather the data points required to produce analytics and reporting needs to be done to aggregate and disseminate the analytics that were produced. This is vital for policy makers to make informed and timely decisions. The Public Health Agency of Canada's stakeholders include the Privy Council, the Chief Public Health Officer, the media, as well as the general public. To meet the needs of around-the-clock surveillance and timely reporting, the Emergency Centre at FAC was activated and various roles were created and staffed using employees from across the agency on a rotation basis. These employees came with various skill sets and most of the processes were initially created using SAS and Excel. These tools required a lot of knowledge transfer, manual labour, and became less effective as data sources increased in number, data sets increased in size. The following slides will outline a few of these inefficiencies. Firstly, the source files for various metrics were large Excel files stored on a network drive. Using SAS, these files took long periods to open and access. In addition, access was often limited due to the fact that SAS would not be able to open these files if another user had them open. Given the time-sensitive nature of reporting needs, this often posed an issue. Large portions of reporting had to be done manually to produce the necessary graphics. Numerous SAS files had to be run separately. The plots that were output had to then be copied and manually placed into a deck. This required careful placement onto slides, ensuring alignments were correct not only on the slide, but across slides. This work is time-consuming and subject to human error. The numeric calculations that were generated had to also be manually input into tables. This required copying and pasting numbers into a table in a slide, then manually formatting the numbers. Formatting included commas, decimals, percentage signs, font, color, among other things. For example, an increase in case counts was manually formatted using red font to indicate a negative trend, while a decrease had to be manually formatted using green to indicate a positive trend. This had to be done for numerous jurisdictions and indicators across multiple decks. The manual nature of the task also made reporting susceptible to human error. Finally, there were issues with version control. Due to the fact that different teams were responsible for their own code, the data the agency was reporting became susceptible to inconsistencies. For example, if manual corrections needed to be made to case data retroactively, it might be applied in one deck but not in another due to communication issues and the use of different codes. The same is true for when formulas had to be updated. In addition, the poor version control made it difficult for new team members to adapt when rotating into the team. To resolve these issues, the innovation team at FAC introduced R to simplify, streamline, and make more efficient our processes. Multiple packages in R are assisted with this effort and the following slides will go through some of them. What initially required the work of multiple epidemiologists was streamlined to one user first through the introduction of R Markdown. R Markdown is a powerful reporting tool that can output tables, objects, plots, and figures into a report with fixed formatting. This is done by running source codes in chunks and using a knitting feature that runs all the source codes and produces an output. The codes that were initially being run using SAS were then rewritten in R and moved to the back end and called upon using R Markdown. This has allowed us to automate our processes that would take multiple hours of labor across several employees to the simple click of a button. Reports are currently being produced using R Markdown in PowerPoint, HTML, PDF, and DocX. Another R package that assisted our effort to improve our reporting was Google Sheets 4. To control for our access issues, the datasets that were being stored in Excel on the network drive were moved to Google Sheets, which is an open-sourced web-based application. Using the Google Sheets package in R, the data was being retrieved with increased speed and without any access issues, as data stored in Google Sheets can be accessed regardless of how many people have it open. To further reduce the burden of manual labor, the data that was being displayed in tables was now being output using the FlexTable package in R. Previously, this data had to be manually entered and formatted. With FlexTable, entire data frames were now being output into documents pre-formatted. Everything from dates, font color, number formats, and footnotes were taken care of using this package. Even dynamic headers were created to incorporate dates. R is also being used to reduce the burden of analysis and commentary. A large amount of time was being devoted nightly to going through the various metrics and analytics and summarizing the findings as well as isolating important data points. This analysis was then placed in bullet points or footnotes into the reports. This is now being automated using simple calculations and base functions like PACE0 in conjunction with R Markdown. Now, providing big picture analysis is automatically done and outliers are automatically detected. With this information, summary sheets are automatically populated. How many provinces reported their data today? Which province saw the largest increase in cases? How about deaths? How has hospitalization changed since the previous day? Packages like LuberDate also proved to be very useful. Some metrics and reporting is only done on a weekly basis. With predefined functions that have the ability to detect weeks based on a given date, producing these analytics became smooth and simple. In addition, utilizing LuberDate functions also allowed for dynamic titles, headers, and footnotes. Finally, to incorporate all the work that was done using R, a package was built for internal use. This unifies and simplifies data intake, data processing, outputs, and upkeep. Each data import, table, figure, and markdown report has been written into its own function. These functions comprise the R package which is utilized by the agency and will continue to expand as new data sources and reports are incorporated. It allows for uniformity in reporting and allows all employees to easily access and generate figures. As our efforts continue, more surveillance programs and products will be incorporated into our package and the use of R across the agency will increase.