The two major recent debris events, the Chinese antisatellite test which destroyed their FY-1C satellite, and the accidental collision of Iridium 33 and Kosmos-2251, are treated separately so that their huge effect can be more readily appreciated.
The data are shown as both individual graphs and as a cumulative sand chart.
The raw data are here
The mass of individual pieces of debris is not known, but is small compared to that of payloads and rocket stages. For example, all of the FY-1C debris totals roughly 1000 kg, the mass of the original satellite. Here is an attempt at estimating the mass of the orbiting satellite population; dry masses are used - i.e. liquid propellant mass is omitted. The y axis of the graph is in (metric) tonnes. The expert reader will spot the Mir and Skylab reentries.
Orbital launches, including Earth orbit and deep space launches. Launch vehicle failures that reached orbit are included; marginal orbit cases are included.
Note that for 2011, US launches exceed those of China only because the Sea Launch partnership is considered to be a US-based launch system.
Below, launch failures are also included, and the data are also shown as percentages.
Number of satellite payloads launched each year on successful orbital launches
Tonnage of satellite payloads launched each year on successful orbital launches
Suborbital launches. The mesospheric launches (green) are mostly weather rockets; NOAA and NASA databases ended in the 1980s so recent decades are missing. The suborbital launches (apogee above 80 km) include science sounding rockets and missile tests. Data on Russian and Chinese missile tests and Scud-class launches in developing-world conflicts are still significantly incomplete. The spike in the 1940s represents WWII V-2 launches with known dates.
Orbital launch attempts. This figure is in 'sand chart' style where the height of the histogram bins is the sum of all the categories, so that the different categories don't overlap each other. Note that it took humanity about 5 years to figure out the orbital launch trick and reduce failures from an initial 50 percent to a few percent. The same data is also shown as a percentage of the total.