History of Django's popularity
What sequence of events made Django the most popular Python web framework .. and still so? Even though several other frameworks exist.
Note: This question is neither argumentative nor confrontational. I merely asked for (objective) "the sequence of events" that lead to its actual popularity. Being aware of the dynamics of software acceptance, I am not intending anyone to get into arguments over technical superiority.
Answer
Many Python web frameworks already existed when Django appeared in 2005 -- indeed, the joke was already going around, by then, that Python is "the language with more web frameworks than keywords" (and Guido rejected my proposal to fix that in Py3k by adding many, many more keywords). Now "django" per se is a bit ambiguous as a search term (it's also the name of a popular guitar player whose life inspired a Woody Allen movie, etc, etc), nevertheless adding "python" to the search to remove those other meanings you can see e.g. in this graph how its relative popularity changed in comparison to another classic Python web framework, Zope. Mostly steady growth quarter over quarter, with a huge surprising jump up at the start of Q2 2008... which just happens to coincide with the date in which Google announced App Engine (it's impossible to prove causation in such case, but the coincidence is at least interesting;-).
App Engine essentially rules out any Python web framework that deeply depends on custom C-coded components, or intrinsically require "heavily relational" functionality; out of those that run well with just pure Python code, Django is probably the one that App Engine most directly and visibly supports. However, this was just a boost, adding to Django's underlying healthy growth trend. The explanation for that trend (and indeed for App Engine team and users' decision to support Django so well) must lie in characteristics that are intrinsic to Django itself.
Django is sometimes criticized (including by... yours truly;-) for being "too magical" or "too monolithic", compared to alternatives such as Pylons, TurboGears, Werkzeug, &c, which are lighter-weight (esp. the latter, my favorite;-), more transparent, and allow easier swapping in and out of specific components (ORM, templating, &c). However, Django's popularity tells us that, for most people interested in developing server-side web sites and apps, these Django design choices are perceived positively: Django's seen as a very rich and well integrated framework (and it does have a lot of add-ons and contributed "plugins", but those are more a consequence than a cause of its ascendancy).
Ease of getting started, automagical "admin pages", and the like -- as well as the fact that Django can be bent to make really rich and complex sites/apps and accommodate peculiar or unique requirements, with much skill and some work -- are probably the "killer features". To use Werkzeug at its best, you need to understand HTTP, and WSGI, and pick and integrate your favorite storage and templating -- developers of Python-based web sites and apps (like, in a sense, users of Rails, or users of even-more-popular PHP!-) are "voting with their mindshare" for an environment in which they need not necessarily do any of that, but can mostly focus on their application domain. I will have to admit that they probably do have a point;-).