An R package to help facilitate talent reviews

Attempting to script my way out of copy+paste gslides.

Talent conversations is something woefully unsupported in terms of technology. The following R package represents a few Sunday evenings work, where I tried to put together a more cohesive and fair way to capture and share talent profiles in our team between managers.

This led to an R package, available at

Table of Contents

The why

The main point of the app is to make the talent summaries more accessible between managers in a way that allows easy maintenance, while ensuring the right data is shown the right way in the various types of talent related interactions we have during the year. This means that rather than browsing through slides (and copy and pasting between years), you gain access to the team info through a highly customisible app.

A key principle of the app is that no data is held in the code - instead it pulls from a gsheet each time the app is rendered.


The app runs off a gsheet, so first step is to make a copy of this gsheet, but replacing the content with real content (and ensuring permissions are restricted!).

Then on the app side, no configuration is needed, instead you just need to tell the app where the gsheet is, using an environment variable called TALENTREVIEW_SHEETID = gsheet_id.

With that environment variable set, you can install the app via remotes::install_github("epijim/TalentFlow"), then run it via golem::run_dev().

The app


The main part of the app is the cards, which displays the team info on bootstrap cards. On all pages, there is also the filter sidepanel, which lets you refine which team members are displayed.

This page is designed for quick checks on the team, but also is the primary tool for talent reviews. The filters on the left help to focus it in (e.g. to talk about people close to promotion).

Job levels

The next screen shows a simple histogram by job level.


Then there is the ratings screen, which has a selection of plots capturing info on the team performance.


The last screen has summary tables across the metrics plotted in early screens.

The package

The package is available at

James Black
James Black
PhD (Cantab)

James Black. Kiwi | Epidemiologist | Data Scientist | Engineering enthusiast.

comments powered by Disqus