Russian tactics have changed dramatically and officers in particular have gotten increasingly more afraid of getting killed by leading from the front and being in HIMARS range. You can literally tell when HIMARS strikes hit officer concentrations (HQ buildings mostly) by the peaks in July and August.

More specifically, for the last few months (at least on some 50km of front line), Russian troops are sent (~10 at a time) on mindless assaults to locate Ukrainian positions and relay the information to artillery. They repeat the operation once artillery need more targets or to validate if the last shelling was effective.

All of this is widely reported by Russian troops on the ground.

The number of officers per day is the left axis, it peaked at 32 on March 6. The right axis, and white line, is tracking the number of troops "eliminated" by the AFU (their definition has never been clarified AFAIK, but it's widely understood not to include wounded).

As for the data, you can download/export the data presented in the visualization by hovering above the graph and 3 dots will appear on the top-right corner. You can export that data from there.

If you want my own data, this is the spreadsheet I keep: https://docs.google.com/spreadsheets/d/1JSPEJpjMMTyAh5NaEVbRB_GEXgrjAFjFpeN3_myLKTc/edit#gid=1132162938

If you want the full breakdown of officer ranks, Twitter user @KilledInUkraine and team is the authority on the subject: https://docs.google.com/spreadsheets/d/1_bpIqkzD88hlSpA-PDZenSQGNnVnxz3lwYHKViSyuUc/edit#gid=1361265165

It's useful because of disinformation and claims to disinformation/propaganda.

The data shows what we should expect, yes, and when copious amount of bullshit is flung all over the place, I find it important to validate the information available.

It's also useful to show the effect of changing Russian and Ukrainian tactics (and to prove that those tactics have a real effect).

The dashboard linked is produced by @ragnarbjartur, but I'm the one who collected the officers's date of death.

Here's a visualization for the first 7 months.

It takes at least 5 months to get 90%+ of the data for a particular time period, by my estimate.

The troops:officer killed ratio was approximately 30:1 during the first month of the invasion and about 50:1 during the 6-7 months that followed. I'm also using known publication date of officer death that's not showing in the dashboard to make those estimates. I suspect the ratio increased to ~70:1 or perhaps even higher after the Kherson evacuation.