Tanay Pratap Singh/ live

Four dashboards · real public data · computed in your browser

Live data, measured as it arrives

Each dashboard connects your browser straight to a public source and does the math on the page, with no server in between and no simulated numbers. Every estimate is checked against something it can be held to: an exact count, the exchange's own checksum, a published law, or a physical model.

The edit stream dashboard: edits flowing past in lanes by wiki, above live counts
connecting to Wikimedia
01 · Wikipedia

Wikipedia, edit by edit

Every change to Wikipedia and its sister sites flows past as it happens. Three memory saving counting methods run beside exact counts, so you can watch how far each one drifts from the truth.

  • Counts distinct editors from 4 KB of memory, next to the exact count (HyperLogLog)
  • Finds the busiest pages, with a guaranteed error on every count (Space-Saving)
  • Flags sudden bursts of editing as they happen
Source Wikimedia EventStreamsArrives as a continuous event stream
The order book dashboard: a heatmap of waiting buy and sell orders with trades drawn over it
connecting to Kraken
02 · Bitcoin order book

A Bitcoin order book you can trust

The page rebuilds Kraken's order book from more than a hundred updates a second and checks it against the exchange's own checksum after every one. Then it measures how buying and selling pressure moves the price.

  • Every update verified, with automatic recovery on a mismatch
  • A heatmap of where buyers and sellers are waiting, with every trade
  • Price impact of order flow and volatility, computed live
Source Kraken public WebSocketArrives as a snapshot, then changes
The earthquake dashboard: a world map of the week's earthquakes colored by depth
checking the USGS feed
03 · Earthquakes

Earthquakes, and how the record changes

The past week of earthquakes from the US Geological Survey, refreshed every minute. The page tests the classic law of how often big quakes happen, fits how aftershocks fade, and logs every time the catalog corrects itself.

  • How much rarer each bigger magnitude is, with an uncertainty range (Gutenberg-Richter)
  • How fast aftershocks die away, fitted from the data (Omori-Utsu)
  • Every magnitude revision and deletion, as it happens
Source USGS earthquake feedsArrives as a poll every minute, merged
The space weather dashboard: solar wind travelling from a spacecraft at L1 toward Earth
checking NOAA
04 · Space weather

The solar wind, on its way to Earth

A spacecraft about 1.5 million km toward the Sun measures the solar wind roughly an hour before it reaches us. The page follows it to Earth, shows how hard it presses on Earth's magnetic field, and maps the aurora against Syracuse.

  • An arrival countdown from the spacecraft's real position
  • The edge of Earth's magnetic field, drawn live (Shue model)
  • X-ray flares, the storm forecast, and the aurora map
Source NOAA Space Weather Prediction CenterArrives as nine feeds, joined on time

How these are built

Your browser connects straight to each public source. Nothing passes through a server of mine and nothing is stored between visits. Data arrives as fast as the source sends it, and the page redraws a few times a second on its own schedule, so a burst of messages never freezes the screen.

The math lives apart from the drawing code and has its own test suite. Each method is checked against something with a known answer: the sketches against their error bounds, the order book against a captured exchange checksum, the earthquake statistics against synthetic catalogs, and the physics against published reference values. The run logs behind every number in the write ups sit in the repository beside the code.

Every chart has a table view. Colors are checked for color blindness, and each feed's status shows as a shape and a word as well as a color. Pausing the display freezes the drawing while the data keeps flowing in, so counts stay correct.

Where a number rests on an assumption, the caveat sits beside it: clock offset in delay figures, mixed magnitude scales in the earthquake statistics, and the simple travel time model behind the solar wind countdown.

Data: Wikimedia EventStreams, Kraken public WebSocket API, U.S. Geological Survey, NOAA Space Weather Prediction Center. None of this is investment advice. Back to the portfolio