brawlerleg/russian-fuel-crisis-2026

This project is an interactive chart, viewable as a website, that lines up two different kinds of data about a fuel shortage in Russia during 2026. One line on the chart shows how often people inside Russia searched Google for the phrase meaning no petrol, week by week. The other layer shows individual drone strikes on oil refineries and fuel terminals, drawn as circles whose size reflects how much fuel that facility could normally process or move each year. Hovering over any circle shows the facility's name, its capacity, and how far it sits from the front line of the conflict. A reader can also move the cursor along the search interest line to read its value on any single day, and can zoom the time window to the last 30, 90, or 180 days, or the full range covered by the data. The page itself is plain HTML, CSS, and JavaScript with no build step or framework, and it works in both Russian and English. Because the page loads its data with a fetch request, it needs to be run through a small local web server rather than opened directly as a file, which the README shows how to do with a one line Python command. Behind the chart, a set of Python scripts rebuild the underlying data: one pulls a table of refinery and terminal strikes from a Wikipedia article and cleans up its inconsistent date formats, another downloads the Google Trends search data, and a third merges both into the single JSON file the webpage reads. The README is careful to flag several limits of the data. A rise in searches after a strike does not prove the strike caused it, since both can move with news cycles. The search line measures attention, not the price people actually paid for fuel. The capacity figures mix refinery processing capacity with much larger terminal transit capacity, so totals should be read with that in mind, and Google Trends values are relative to the chosen window rather than absolute counts. The code is released under the MIT license, while the strike data itself comes from Wikipedia under a separate Creative Commons license that requires attribution.