There are numerous issues with the statement in the OP.
The New York Times is getting its information from the article http://www.newamerica.org/in-depth/terrorism-in-america/what-threat-united-states-today/#americas-layered-defenses
That source leaves out the following fatal, confirmed terrorism, US incidents found in the terrorism database (as well as other incidents flagged as "possible" terrorism):
07/23/2015 shooting by John Russell Houser at theater
12/20/2014 shooting by Ismaaiyl Brinsley
12/18/2014 shooting by Justin Nojan Sullivan "so he could purchase a rifle to carry out an Islamic State of Iraq and the Levant (ISIL) inspired attack"
11/01/2013 shooting by Paul Ciancia at Los Angeles International Airport
(also, if the NYT phrase "after 9/11" on the graph is taken literally, the anthrax attacks are omitted)
On the other hand, the "new america" source adds incidents that are not considered terrorism (or even possible terrorism) by the database, specifically:
-a 2012 shooting of police
-a 2011 murder by Isaac Aguigui and others of his wife and unborn child and two others
-a 2011 murder by David Pedersen and Holly Grigsby of Pedersen's father and stepmother and two others
-a 2010 murder at a Carlisle, PA gun range
-a 2009 shooting of police officers when they responded to a domestic dispute call
-a 2009 murder in Brockten, MA
-a 2004 Tulsa, OK bank robbery
The "new america" article originally just graphed "right wing" and "jihadist" attacks. The New York Times relabeled the ones that were supposedly "right wing" as "Non-Islamic" and the ones labeled "jihadist" as "Islamic".
All the incidents that "new america" decided to add were labeled as "right wing". Of the 18 incidents that "new america" labeled "right wing", 7 were not in the terrorism database.
Other factors to consider are:
The 25 factor is based upon fatalities, not incidents or perpetrators.
Sometimes people have come to the USA to conduct attacks, not all attacks are internal. Just using the demographics of the USA is not valid considering that some attacks are by external persons.
The data are cherry-picked (not necessarily intentionally) timewise to only include the period after the 9/11/01 attacks and before the 2016 Orlando attack.
The statement in the OP article is:
Muslims make up only about 2% of the population of America. If they are killing half the victims of terrorism, that means that a Muslim perpetrator is 25 times more likely to kill someone in a terrorist attack than a non-Muslim.
This statement contains a math error. If 2% of the population killed an equal number of victims as the remaining 98%, they killed at 98/2 = 49 times the rate.
- The most recent estimate by Pew Research is that there are 3.3 million Muslims out of 322 million total people in the USA, or 1.0%.